November 2009 - galaxy-dev - lists.galaxyproject.org

[hg] galaxy 3082: Bug fixes and minor updates.
by Greg Von Kuster 23 Nov '09

23 Nov '09

details: http://www.bx.psu.edu/hg/galaxy/rev/70ae74578254 changeset: 3082:70ae74578254 user: Enis Afgan <afgane(a)gmail.com> date: Wed Nov 04 17:19:18 2009 -0500 description: Bug fixes and minor updates. diffstat: lib/galaxy/cloud/providers/ec2.py | 20 ++++- lib/galaxy/cloud/providers/eucalyptus.py | 36 ++++++--- lib/galaxy/web/controllers/cloud.py | 122 ++++++++++-------------------- templates/cloud/configure_cloud.mako | 20 +++- 4 files changed, 96 insertions(+), 102 deletions(-) diffs (503 lines): diff -r 89276f68513a -r 70ae74578254 lib/galaxy/cloud/providers/ec2.py --- a/lib/galaxy/cloud/providers/ec2.py Tue Nov 03 13:20:26 2009 -0500 +++ b/lib/galaxy/cloud/providers/ec2.py Wed Nov 04 17:19:18 2009 -0500 @@ -45,7 +45,8 @@ store_states = Bunch( IN_USE = "in-use", - CREATING = "creating" + CREATING = "creating", + ERROR = "error" ) class EC2CloudProvider( object ): @@ -94,7 +95,7 @@ """ Establishes EC2 cloud connection using user's credentials associated with given UCI """ - log.debug( '##### Establishing EC2 cloud connection' ) + log.debug( 'Establishing %s cloud connection' % self.type ) provider = uci_wrapper.get_provider() try: region = RegionInfo( None, provider.region_name, provider.region_endpoint ) @@ -276,7 +277,7 @@ if uci_wrapper.get_state() != uci_states.ERROR: # Start an instance - log.debug( "***** Starting instance for UCI '%s'" % uci_wrapper.get_name() ) + log.debug( "Starting instance for UCI '%s'" % uci_wrapper.get_name() ) #TODO: Once multiple volumes can be attached to a single instance, update 'userdata' composition userdata = uci_wrapper.get_store_volume_id()+"|"+uci_wrapper.get_access_key()+"|"+uci_wrapper.get_secret_key() log.debug( 'Using following command: conn.run_instances( image_id=%s, key_name=%s, security_groups=[%s], user_data=[OMITTED], instance_type=%s, placement=%s )' @@ -401,9 +402,19 @@ model.CloudStore.c.status==store_states.CREATING, model.CloudStore.c.status==None ) ).all() for store in stores: - if self.type == store.uci.credentials.provider.type: + if self.type == store.uci.credentials.provider.type and store.volume_id != None: log.debug( "[%s] Running general status update on store '%s'" % ( store.uci.credentials.provider.type, store.volume_id ) ) self.updateStore( store ) + else: + log.error( "[%s] There exists an entry for UCI (%s) storage volume without an ID. Storage volume might have been created with " + "cloud provider though. Manual check is recommended." % ( store.uci.credentials.provider.type, store.uci.name ) ) + store.uci.error = "There exists an entry in local database for a storage volume without an ID. Storage volume might have been created " \ + "with cloud provider though. Manual check is recommended. After understanding what happened, local database etry for given " \ + "storage volume should be updated." + store.status = store_states.ERROR + store.uci.state = uci_states.ERROR + store.uci.flush() + store.flush() # Attempt at updating any zombie UCIs (i.e., instances that have been in SUBMITTED state for longer than expected - see below for exact time) zombies = model.UCI.filter_by( state=uci_states.SUBMITTED ).all() @@ -605,6 +616,7 @@ Establishes and returns connection to cloud provider. Information needed to do so is obtained directly from uci database object. """ + log.debug( 'Establishing %s cloud connection' % self.type ) a_key = uci.credentials.access_key s_key = uci.credentials.secret_key # Get connection diff -r 89276f68513a -r 70ae74578254 lib/galaxy/cloud/providers/eucalyptus.py --- a/lib/galaxy/cloud/providers/eucalyptus.py Tue Nov 03 13:20:26 2009 -0500 +++ b/lib/galaxy/cloud/providers/eucalyptus.py Wed Nov 04 17:19:18 2009 -0500 @@ -46,7 +46,8 @@ store_states = Bunch( IN_USE = "in-use", - CREATING = "creating" + CREATING = "creating", + ERROR = "error" ) class EucalyptusCloudProvider( object ): @@ -96,7 +97,7 @@ """ Establishes eucalyptus cloud connection using user's credentials associated with given UCI """ - log.debug( '##### Establishing eucalyptus cloud connection' ) + log.debug( 'Establishing %s cloud connection.' % self.type ) provider = uci_wrapper.get_provider() try: euca_region = RegionInfo( None, provider.region_name, provider.region_endpoint ) @@ -170,14 +171,14 @@ and registers relevant information in Galaxy database. """ conn = self.get_connection( uci_wrapper ) + + # Because only 1 storage volume may be created at UCI config time, index of this storage volume in local Galaxy DB w.r.t + # current UCI is 0; therefore, it can be referenced in following code + log.info( "Creating volume in zone '%s'..." % uci_wrapper.get_uci_availability_zone() ) if uci_wrapper.get_uci_availability_zone()=='': log.info( "Availability zone for UCI (i.e., storage volume) was not selected, using default zone: %s" % self.zone ) uci_wrapper.set_store_availability_zone( self.zone ) - #TODO: check if volume associated with UCI already exists (if server crashed for example) and don't recreate it - log.info( "Creating volume in zone '%s'..." % uci_wrapper.get_uci_availability_zone() ) - # Because only 1 storage volume may be created at UCI config time, index of this storage volume in local Galaxy DB w.r.t - # current UCI is 0, so reference it in following methods vol = conn.create_volume( uci_wrapper.get_store_size( 0 ), uci_wrapper.get_uci_availability_zone(), snapshot=None ) uci_wrapper.set_store_volume_id( 0, vol.id ) @@ -209,15 +210,14 @@ failedList.append( v.volume_id ) # Delete UCI if all of associated - log.debug( "count=%s, len(vl)=%s" % (count, len( vl ) ) ) if count == len( vl ): uci_wrapper.delete() else: log.error( "Deleting following volume(s) failed: %s. However, these volumes were successfully deleted: %s. \ - MANUAL intervention and processing needed." % ( failedList, deletedList ) ) + Manual intervention and processing needed." % ( str( failedList ), str( deletedList ) ) ) uci_wrapper.change_state( uci_state=uci_states.ERROR ) - uci_wrapper.set_error( "Deleting following volume(s) failed: "+failedList+". However, these volumes were successfully deleted: "+deletedList+". \ - MANUAL intervention and processing needed." ) + uci_wrapper.set_error( "Deleting following volume(s) failed: "+str(failedList)+". However, these volumes were \ + successfully deleted: "+str(deletedList)+". Manual intervention and processing needed." ) def addStorageToUCI( self, name ): """ Adds more storage to specified UCI """ @@ -250,7 +250,7 @@ uci_wrapper.set_mi( i_index, mi_id ) if uci_wrapper.get_state() != uci_states.ERROR: - log.debug( "***** Starting UCI instance '%s'" % uci_wrapper.get_name() ) + log.debug( "Starting UCI instance '%s'" % uci_wrapper.get_name() ) log.debug( 'Using following command: conn.run_instances( image_id=%s, key_name=%s )' % ( mi_id, uci_wrapper.get_key_pair_name( i_index ) ) ) try: reservation = conn.run_instances( image_id=mi_id, key_name=uci_wrapper.get_key_pair_name( i_index ) ) @@ -368,9 +368,19 @@ model.CloudStore.c.status==store_states.CREATING, model.CloudStore.c.status==None ) ).all() for store in stores: - if self.type == store.uci.credentials.provider.type: + if self.type == store.uci.credentials.provider.type and store.volume_id != None: log.debug( "[%s] Running general status update on store '%s'" % ( store.uci.credentials.provider.type, store.volume_id ) ) self.updateStore( store ) + else: + log.error( "[%s] There exists an entry for UCI (%s) storage volume without an ID. Storage volume might have been created with " + "cloud provider though. Manual check is recommended." % ( store.uci.credentials.provider.type, store.uci.name ) ) + store.uci.error = "There exists an entry in local database for a storage volume without an ID. Storage volume might have been created " \ + "with cloud provider though. Manual check is recommended. After understanding what happened, local database etry for given " \ + "storage volume should be updated." + store.status = store_states.ERROR + store.uci.state = uci_states.ERROR + store.uci.flush() + store.flush() # Attempt at updating any zombie UCIs (i.e., instances that have been in SUBMITTED state for longer than expected - see below for exact time) zombies = model.UCI.filter_by( state=uci_states.SUBMITTED ).all() @@ -455,6 +465,7 @@ conn = self.get_connection_from_uci( uci ) try: + log.debug( "vol id: " % store.volume_id ) vl = conn.get_all_volumes( [store.volume_id] ) except boto.exception.EC2ResponseError, e: log.error( "Retrieving volume(s) from cloud for UCI '%s' failed: " % ( uci.name, str(e) ) ) @@ -570,6 +581,7 @@ Establishes and returns connection to cloud provider. Information needed to do so is obtained directly from uci database object. """ + log.debug( 'Establishing %s cloud connection.' % self.type ) a_key = uci.credentials.access_key s_key = uci.credentials.secret_key # Get connection diff -r 89276f68513a -r 70ae74578254 lib/galaxy/web/controllers/cloud.py --- a/lib/galaxy/web/controllers/cloud.py Tue Nov 03 13:20:26 2009 -0500 +++ b/lib/galaxy/web/controllers/cloud.py Wed Nov 04 17:19:18 2009 -0500 @@ -52,11 +52,14 @@ SHUTTING_DOWN = "shutting-down" ) +store_states = Bunch( + IN_USE = "in-use", + CREATING = "creating", + ERROR = "error" +) + class CloudController( BaseController ): - -# def __init__( self ): -# self.cloudManager = CloudManager() - + @web.expose def index( self, trans ): return trans.fill_template( "cloud/index.mako" ) @@ -76,22 +79,22 @@ liveInstances = trans.sa_session.query( model.UCI ) \ .filter_by( user=user ) \ - .filter( or_( model.UCI.c.state==uci_states.RUNNING, #"running", - model.UCI.c.state==uci_states.PENDING, #"pending", - model.UCI.c.state==uci_states.SUBMITTED, #"submitted", - model.UCI.c.state==uci_states.SUBMITTED_UCI, #"submittedUCI", - model.UCI.c.state==uci_states.SHUTTING_DOWN, #"shutting-down", + .filter( or_( model.UCI.c.state==uci_states.RUNNING, + model.UCI.c.state==uci_states.PENDING, + model.UCI.c.state==uci_states.SUBMITTED, + model.UCI.c.state==uci_states.SUBMITTED_UCI, + model.UCI.c.state==uci_states.SHUTTING_DOWN, model.UCI.c.state==uci_states.SHUTTING_DOWN_UCI ) ) \ .order_by( desc( model.UCI.c.update_time ) ) \ .all() prevInstances = trans.sa_session.query( model.UCI ) \ .filter_by( user=user ) \ - .filter( or_( model.UCI.c.state==uci_states.AVAILABLE, #"available", - model.UCI.c.state==uci_states.NEW, #"new", - model.UCI.c.state==uci_states.NEW_UCI, #"newUCI", - model.UCI.c.state==uci_states.ERROR, #"error", - model.UCI.c.state==uci_states.DELETING, #"deleting", + .filter( or_( model.UCI.c.state==uci_states.AVAILABLE, + model.UCI.c.state==uci_states.NEW, + model.UCI.c.state==uci_states.NEW_UCI, + model.UCI.c.state==uci_states.ERROR, + model.UCI.c.state==uci_states.DELETING, model.UCI.c.state==uci_states.DELETING_UCI ) ) \ .order_by( desc( model.UCI.c.update_time ) ) \ .all() @@ -99,8 +102,8 @@ # Check after update there are instances in pending state; if so, display message pendingInstances = trans.sa_session.query( model.UCI ) \ .filter_by( user=user ) \ - .filter( or_( model.UCI.c.state==uci_states.PENDING, #"pending" , \ - model.UCI.c.state==uci_states.SUBMITTED, #"submitted" , \ + .filter( or_( model.UCI.c.state==uci_states.PENDING, + model.UCI.c.state==uci_states.SUBMITTED, model.UCI.c.state==uci_states.SUBMITTED_UCI ) ) \ .all() if pendingInstances: @@ -124,6 +127,7 @@ def makeDefault( self, trans, id=None ): """ Set current credentials as default. + *NOT USED* """ currentDefault = get_default_credentials (trans) if currentDefault: @@ -137,8 +141,7 @@ # TODO: Fix bug that when this function returns, top Galaxy tab bar is missing from the webpage return self.list( trans ) #trans.fill_template( "cloud/configure_cloud.mako", #awsCredentials = awsCredentials ) - - + @web.expose @web.require_login( "start Galaxy cloud instance" ) def start( self, trans, id, type='m1.small' ): @@ -149,19 +152,9 @@ uci = get_uci( trans, id ) mi = get_mi( trans, uci, type ) stores = get_stores( trans, uci ) - # Ensure instance is not already running (or related state) and store relevant data + # Ensure instance is available and then store relevant data # into DB to initiate instance startup by cloud manager - if ( len(stores) is not 0 ) and \ - ( uci.state != uci_states.SUBMITTED ) and \ - ( uci.state != uci_states.SUBMITTED_UCI ) and \ - ( uci.state != uci_states.PENDING ) and \ - ( uci.state != uci_states.DELETING ) and \ - ( uci.state != uci_states.DELETING_UCI ) and \ - ( uci.state != uci_states.DELETED ) and \ - ( uci.state != uci_states.RUNNING ) and \ - ( uci.state != uci_states.NEW_UCI ) and \ - ( uci.state != uci_states.NEW ) and \ - ( uci.state != uci_states.ERROR ): + if ( len(stores) is not 0 ) and ( uci.state == uci_states.AVAILABLE ): instance = model.CloudInstance() instance.user = user instance.image = mi @@ -181,7 +174,10 @@ "instance description." ) return self.list( trans ) - error( "Cannot start instance that is in state '%s'." % uci.state ) + if len(stores) == 0: + error( "This instance does not have any storage volumes associated it and thus cannot be started." ) + else: + error( "Cannot start instance that is in state '%s'." % uci.state ) return self.list( trans ) @web.expose @@ -228,7 +224,10 @@ trans.set_message( "Galaxy instance '%s' marked for deletion." % name ) return self.list( trans ) - trans.set_message( "Instance '%s' is already marked for deletion." % uci.name ) + if uci.state != uci_states.ERROR: + trans.set_message( "Cannot delete instance in state ERROR." ) + else: + trans.set_message( "Instance '%s' is already marked for deletion." % uci.name ) return self.list( trans ) @web.expose @@ -251,8 +250,6 @@ .filter_by( user=user, state=instance_states.TERMINATED, uci_id=id ) \ .order_by( desc( model.CloudInstance.c.update_time ) ) \ .all() - - log.debug( "id: %s" % id ) return trans.fill_template( "cloud/view_usage.mako", prevInstances = prevInstances ) @@ -284,6 +281,12 @@ providersToZones[storedCred.name] = ['Unknown provider zone'] if instanceName: + # Check if volume size is entered as an integer + try: + volSize = int( volSize ) + except ValueError: + error['vol_error'] = "Volume size must be integer value between 1 and 1000." + # Create new user configured instance try: if trans.app.model.UCI \ @@ -299,9 +302,6 @@ error['vol_error'] = "You must specify volume size as an integer value between 1 and 1000." elif ( int( volSize ) < 1 ) or ( int( volSize ) > 1000 ): error['vol_error'] = "Volume size must be integer value between 1 and 1000." -# elif type( volSize ) != type( 1 ): # Check if volSize is int -# log.debug( "volSize='%s'" % volSize ) -# error['vol_error'] = "Volume size must be integer value between 1 and 1000." elif zone=='': error['zone_error'] = "You must select zone where this UCI will be registered." else: @@ -311,14 +311,14 @@ uci.credentials = trans.app.model.CloudUserCredentials.filter( trans.app.model.CloudUserCredentials.table.c.name==credName ).first() uci.user= user - uci.total_size = volSize # This is OK now because new instance is being created. + uci.total_size = volSize # This is OK now because new instance is being created and only one storage volume can be created at UCI creation time uci.state = uci_states.NEW_UCI storage = model.CloudStore() storage.user = user storage.uci = uci storage.size = volSize - storage.availability_zone = zone # TODO: Give user choice here. Also, enable region selection. + storage.availability_zone = zone # Persist session = trans.sa_session session.save_or_update( uci ) @@ -333,7 +333,7 @@ except AttributeError, ae: inst_error = "No registered cloud images. You must contact administrator to add some before proceeding." log.debug("AttributeError: %s " % str( ae ) ) - + return trans.fill_template( "cloud/configure_uci.mako", instanceName = instanceName, credName = storedCreds, @@ -341,14 +341,7 @@ zone = zone, error = error, providersToZones = providersToZones ) - - return trans.show_form( - web.FormBuilder( web.url_for(), "Configure new instance", submit_text="Add" ) - .add_text( "instanceName", "Instance name", value="Unnamed instance", error=inst_error ) - .add_text( "credName", "Name of registered credentials to use", value="", error=cred_error ) - .add_text( "volSize", "Permanent storage size (1GB - 1000GB)" - "<br />Note: you will be able to add more storage later", value='', error=vol_error ) ) - + @web.expose @web.require_admin def addNewImage( self, trans, image_id='', manifest='', state=None ): @@ -851,6 +844,7 @@ user = trans.get_user() stores = trans.sa_session.query( model.CloudStore ) \ .filter_by( user=user, uci_id=uci.id ) \ + .filter( model.CloudStore.c.status != store_states.ERROR ) \ .all() return stores @@ -868,19 +862,6 @@ return instances -def get_cloud_instance( conn, instance_id ): - """ - Returns a cloud instance representation of the instance id, i.e., cloud instance object that cloud API can be invoked on - """ - # get_all_instances func. takes a list of desired instance id's, so create a list first - idLst = list() - idLst.append( instance_id ) - # Retrieve cloud instance based on passed instance id. get_all_instances( idLst ) method returns reservation ID. Because - # we are passing only 1 ID, we can retrieve only the first element of the returning list. Furthermore, because (for now!) - # only 1 instance corresponds each individual reservation, grab only the first element of the returned list of instances. - cloudInstance = conn.get_all_instances( [instance_id] )[0].instances[0] - return cloudInstance - def get_connection( trans, credName ): """ Establishes EC2 connection using user's default credentials @@ -900,22 +881,3 @@ else: error( "You must specify default credentials before starting an instance." ) return 0 - -def get_keypair_name( trans ): - """ - Generate keypair using user's default credentials - """ - conn = get_connection( trans ) - - log.debug( "Getting user's keypair" ) - key_pair = conn.get_key_pair( 'galaxy-keypair' ) - - try: - return key_pair.name - except AttributeError: # No keypair under this name exists so create it - log.debug( 'No keypair found, creating keypair' ) - key_pair = conn.create_key_pair( 'galaxy-keypair' ) - # TODO: Store key_pair.material into instance table - this is the only time private key can be retrieved - # Actually, probably return key_pair to calling method and store name & key from there... - - return key_pair.name diff -r 89276f68513a -r 70ae74578254 templates/cloud/configure_cloud.mako --- a/templates/cloud/configure_cloud.mako Tue Nov 03 13:20:26 2009 -0500 +++ b/templates/cloud/configure_cloud.mako Wed Nov 04 17:19:18 2009 -0500 @@ -26,19 +26,24 @@ var elem = '#' + data[i].id; // Because of different list managing 'live' vs. 'available' instances, reload url on various state changes old_state = $(elem + "-state").text(); + prev_old_state = $(elem + "-state-p").text(); new_state = data[i].state; //console.log( "old_state[%d] = %s", i, old_state ); + //console.log( "prev_old_state[%d] = %s", i, prev_old_state ); //console.log( "new_state[%d] = %s", i, new_state ); if ( ( old_state=='pending' && new_state=='running' ) || ( old_state=='shutting-down' && new_state=='available' ) || \ ( old_state=='running' && new_state=='available' ) || ( old_state=='running' && new_state=='error' ) || \ ( old_state=='pending' && new_state=='error' ) || ( old_state=='pending' && new_state=='available' ) || \ - ( old_state=='submitted' && new_state=='available' ) ) { + ( old_state=='submitted' && new_state=='available' ) || ( prev_old_state.match('newUCI') && new_state=='available' ) || \ + ( prev_old_state.match('new') && new_state=='available' ) ) { var url = "${h.url_for( controller='cloud', action='list')}"; location.replace( url ); } else if ( ( old_state=='running' && new_state=='error' ) || ( old_state=='pending' && new_state=='error' ) || \ ( old_state=='submitted' && new_state=='error' ) || ( old_state=='submittedUCI' && new_state=='error' ) || \ - ( old_state=='shutting-down' && new_state=='error' ) ) { + ( old_state=='shutting-down' && new_state=='error' ) || ( prev_old_state.match('newUCI') && new_state=='error' ) || \ + ( prev_old_state.match('new') && new_state=='error' ) || ( prev_old_state.match('available') && new_state=='error' ) || \ + ( prev_old_state.match('deleting') && new_state=='error' ) || ( prev_old_state.match('deletingUCI') && new_state=='error' ) ) { var url = "${h.url_for( controller='cloud', action='list')}"; location.replace( url ); } @@ -56,13 +61,13 @@ if( !dns ) { $(uci+"-link").text( 'Galaxy starting...' ); // http://stackoverflow.com/questions/275931/how-do-you-make-an-element-flash-… - $(uci+"-link").stop().animate({ fontSize: "14px" }, 1000).animate({ fontSize: "12px" }, 1000); + //$(uci+"-link").stop().animate({ fontSize: "14px" }, 1000).animate({ fontSize: "12px" }, 1000); } else { $(uci+"-link").html( '<div align="right"><a class="action-button" href="http://'+dns+'" target="_blank">' + '<span>Access Galaxy</span>'+ '<img src="' + "${h.url_for( '/static/images/silk/resultset_next.png' )}" + '" /></div>' ); - $(uci+"-link").stop().animate({ fontSize: "14px" }, 1000).animate({ fontSize: "12px" }, 1000); + //$(uci+"-link").stop().animate({ fontSize: "14px" }, 1000).animate({ fontSize: "12px" }, 1000); } } }); @@ -92,6 +97,7 @@ %if cloudCredentials: ## Manage user credentials + <h3>Your registered credentials</h3> <ul class="manage-table-actions"> <li> <a class="action-button" href="${h.url_for( action='add' )}"> @@ -131,7 +137,7 @@ ## ***************************************************** ## Manage live instances <p /> - <h2>Manage your cloud instances</h2> + <h3>Manage your cloud instances</h3> <ul class="manage-table-actions"> <li> <a class="action-button" href="${h.url_for( action='configureNew' )}"> @@ -232,7 +238,7 @@ <a id="pi-${i}-popup" class="popup-arrow" style="display: none;">▼</a> </td> <td>${str(prevInstance.total_size)}</td> - <td> + <td id="${ prevInstance.id }-state-p"> <%state = str(prevInstance.state)%> %if state =='error': <div id="${prevInstance.name}-short"> @@ -249,7 +255,9 @@ error:</a><br /> ${str(prevInstance.error)} <p /> + <div style="font-size:10px;"> <a href="${h.url_for( action='set_uci_state', id=trans.security.encode_id(prevInstance.id), state='available' )}">reset state</a> + </div> </div> %else: ${str(prevInstance.state)}

1 0

[hg] galaxy 3078: Added good bit of exception try/catch blocks i...
by Greg Von Kuster 23 Nov '09

23 Nov '09

details: http://www.bx.psu.edu/hg/galaxy/rev/e19eef93584f changeset: 3078:e19eef93584f user: Enis Afgan <afgane(a)gmail.com> date: Thu Oct 29 17:40:31 2009 -0400 description: Added good bit of exception try/catch blocks in providers' code. Started on addressing zombie UCIs. diffstat: lib/galaxy/cloud/__init__.py | 9 +- lib/galaxy/cloud/providers/ec2.py | 272 +++++++++++++++++++++------------- lib/galaxy/cloud/providers/eucalyptus.py | 283 ++++++++++++++++++++++------------ lib/galaxy/web/controllers/cloud.py | 28 +-- templates/cloud/add_provider.mako | 3 - templates/cloud/configure_cloud.mako | 9 +- templates/cloud/configure_uci.mako | 4 +- templates/cloud/viewInstance.mako | 7 +- 8 files changed, 382 insertions(+), 233 deletions(-) diffs (918 lines): diff -r be631ed97541 -r e19eef93584f lib/galaxy/cloud/__init__.py --- a/lib/galaxy/cloud/__init__.py Thu Oct 29 11:46:43 2009 -0400 +++ b/lib/galaxy/cloud/__init__.py Thu Oct 29 17:40:31 2009 -0400 @@ -34,8 +34,6 @@ CREATING = "creating" ) -JOB_WAIT, JOB_ERROR, JOB_INPUT_ERROR, JOB_INPUT_DELETED, JOB_OK, JOB_READY, JOB_DELETED, JOB_ADMIN_DELETED = 'wait', 'error', 'input_error', 'input_deleted', 'ok', 'ready', 'deleted', 'admin_deleted' - class CloudManager( object ): """ Highest level interface to cloud management. @@ -318,13 +316,16 @@ vol.i_id = instance_id vol.flush() - def set_error( self, error ): + def set_error( self, error, set_state=False ): """ - Sets error field of given UCI in local Galaxy database + Sets error field of given UCI in local Galaxy database. If set_state is set to 'true', + method also sets state of give UCI to 'error' """ uci = model.UCI.get( self.uci_id ) uci.refresh() uci.error = error + if set_state: + uci.state = uci_states.ERROR uci.flush() # --------- Getter methods ----------------- diff -r be631ed97541 -r e19eef93584f lib/galaxy/cloud/providers/ec2.py --- a/lib/galaxy/cloud/providers/ec2.py Thu Oct 29 11:46:43 2009 -0400 +++ b/lib/galaxy/cloud/providers/ec2.py Thu Oct 29 17:40:31 2009 -0400 @@ -12,6 +12,7 @@ import galaxy.eggs galaxy.eggs.require("boto") from boto.ec2.connection import EC2Connection +from boto.ec2.regioninfo import RegionInfo import boto.exception import logging @@ -93,12 +94,23 @@ """ log.debug( '##### Establishing EC2 cloud connection' ) provider = uci_wrapper.get_provider() - region = RegionInfo( None, provider.region_name, provider.region_endpoint ) - conn = EC2Connection( aws_access_key_id=uci_wrapper.get_access_key(), - aws_secret_access_key=uci_wrapper.get_secret_key(), - is_secure=provider.is_secure, - region=region, - path=provider.path ) + try: + region = RegionInfo( None, provider.region_name, provider.region_endpoint ) + except Exception, e: + log.error( "Selecting region with cloud provider failed: %s" % str(e) ) + uci_wrapper.set_error( "Selecting region with cloud provider failed: " + str(e), True ) + return None + try: + conn = EC2Connection( aws_access_key_id=uci_wrapper.get_access_key(), + aws_secret_access_key=uci_wrapper.get_secret_key(), + is_secure=provider.is_secure, + region=region, + path=provider.path ) + except Exception, e: + log.error( "Establishing connection with cloud failed: %s" % str(e) ) + uci_wrapper.set_error( "Establishing connection with cloud failed: " + str(e), True ) + return None + return conn def set_keypair( self, uci_wrapper, conn ): @@ -120,7 +132,7 @@ uci_wrapper.set_key_pair( inst, kp.name, kp.material ) else: log.error( "EC2 response error: '%s'" % e ) - uci_wrapper.set_error( "EC2 response error while creating key pair: " + e ) + uci_wrapper.set_error( "EC2 response error while creating key pair: " + str(e), True ) return kp.name @@ -235,53 +247,65 @@ Starts instance(s) of given UCI on the cloud. """ conn = self.get_connection( uci_wrapper ) -# - self.set_keypair( uci_wrapper, conn ) - i_indexes = uci_wrapper.get_instances_indexes() # Get indexes of i_indexes associated with this UCI whose state is 'None' - log.debug( "Starting instances with IDs: '%s' associated with UCI '%s' " % ( uci_wrapper.get_name(), i_indexes ) ) - for i_index in i_indexes: - mi_id = self.get_mi_id( uci_wrapper.get_type( i_index ) ) - uci_wrapper.set_mi( i_index, mi_id ) + if uci_wrapper.get_state() != uci_states.ERROR: + self.set_keypair( uci_wrapper, conn ) + i_indexes = uci_wrapper.get_instances_indexes() # Get indexes of i_indexes associated with this UCI whose state is 'None' + log.debug( "Starting instances with IDs: '%s' associated with UCI '%s' " % ( uci_wrapper.get_name(), i_indexes ) ) - # Check if galaxy security group exists (and create it if it does not) - security_group = 'galaxyWeb' - log.debug( "Setting up '%s' security group." % security_group ) - sgs = conn.get_all_security_groups() # security groups - gsgt = False # galaxy security group test - for sg in sgs: - if sg.name == security_group: - gsgt = True - # If security group does not exist, create it - if not gsgt: - gSecurityGroup = conn.create_security_group(security_group, 'Security group for Galaxy.') - gSecurityGroup.authorize( 'tcp', 80, 80, '0.0.0.0/0' ) # Open HTTP port - gSecurityGroup.authorize( 'tcp', 22, 22, '0.0.0.0/0' ) # Open SSH port - # Start an instance - log.debug( "***** Starting instance for UCI '%s'" % uci_wrapper.get_name() ) - #TODO: Once multiple volumes can be attached to a single instance, update 'userdata' composition - userdata = uci_wrapper.get_store_volume_id()+"|"+uci_wrapper.get_access_key()+"|"+uci_wrapper.get_secret_key() - log.debug( 'Using following command: conn.run_instances( image_id=%s, key_name=%s, security_groups=[%s], user_data=[OMITTED], instance_type=%s, placement=%s )' - % ( mi_id, uci_wrapper.get_key_pair_name( i_index ), [security_group], uci_wrapper.get_type( i_index ), uci_wrapper.get_uci_availability_zone() ) ) - reservation = conn.run_instances( image_id=mi_id, - key_name=uci_wrapper.get_key_pair_name( i_index ), - security_groups=[security_group], - user_data=userdata, - instance_type=uci_wrapper.get_type( i_index ), - placement=uci_wrapper.get_uci_availability_zone() ) - # Record newly available instance data into local Galaxy database - l_time = datetime.utcnow() - uci_wrapper.set_launch_time( l_time, i_index=i_index ) # format_time( reservation.i_indexes[0].launch_time ) ) - if not uci_wrapper.uci_launch_time_set(): - uci_wrapper.set_uci_launch_time( l_time ) - uci_wrapper.set_reservation_id( i_index, str( reservation ).split(":")[1] ) - # TODO: if more than a single instance will be started through single reservation, change this reference to element [0] - i_id = str( reservation.instances[0]).split(":")[1] - uci_wrapper.set_instance_id( i_index, i_id ) - s = reservation.instances[0].state - uci_wrapper.change_state( s, i_id, s ) - log.debug( "Instance of UCI '%s' started, current state: '%s'" % ( uci_wrapper.get_name(), uci_wrapper.get_state() ) ) - + if uci_wrapper.get_state() != uci_states.ERROR: + for i_index in i_indexes: + mi_id = self.get_mi_id( uci_wrapper.get_type( i_index ) ) + uci_wrapper.set_mi( i_index, mi_id ) + + # Check if galaxy security group exists (and create it if it does not) + security_group = 'galaxyWeb' + log.debug( "Setting up '%s' security group." % security_group ) + sgs = conn.get_all_security_groups() # security groups + gsgt = False # galaxy security group test + for sg in sgs: + if sg.name == security_group: + gsgt = True + # If security group does not exist, create it + if not gsgt: + gSecurityGroup = conn.create_security_group(security_group, 'Security group for Galaxy.') + gSecurityGroup.authorize( 'tcp', 80, 80, '0.0.0.0/0' ) # Open HTTP port + gSecurityGroup.authorize( 'tcp', 22, 22, '0.0.0.0/0' ) # Open SSH port + + if uci_wrapper.get_state() != uci_states.ERROR: + # Start an instance + log.debug( "***** Starting instance for UCI '%s'" % uci_wrapper.get_name() ) + #TODO: Once multiple volumes can be attached to a single instance, update 'userdata' composition + userdata = uci_wrapper.get_store_volume_id()+"|"+uci_wrapper.get_access_key()+"|"+uci_wrapper.get_secret_key() + log.debug( 'Using following command: conn.run_instances( image_id=%s, key_name=%s, security_groups=[%s], user_data=[OMITTED], instance_type=%s, placement=%s )' + % ( mi_id, uci_wrapper.get_key_pair_name( i_index ), [security_group], uci_wrapper.get_type( i_index ), uci_wrapper.get_uci_availability_zone() ) ) + try: + reservation = conn.run_instances( image_id=mi_id, + key_name=uci_wrapper.get_key_pair_name( i_index ), + security_groups=[security_group], + user_data=userdata, + instance_type=uci_wrapper.get_type( i_index ), + placement=uci_wrapper.get_uci_availability_zone() ) + except boto.exception.EC2ResponseError, e: + log.error( "EC2 response error when starting UCI '%s': '%s'" % ( uci_wrapper.get_name(), str(e) ) ) + uci_wrapper.set_error( "EC2 response error when starting: " + str(e), True ) + # Record newly available instance data into local Galaxy database + l_time = datetime.utcnow() + uci_wrapper.set_launch_time( l_time, i_index=i_index ) # format_time( reservation.i_indexes[0].launch_time ) ) + if not uci_wrapper.uci_launch_time_set(): + uci_wrapper.set_uci_launch_time( l_time ) + try: + uci_wrapper.set_reservation_id( i_index, str( reservation ).split(":")[1] ) + # TODO: if more than a single instance will be started through single reservation, change this reference to element [0] + i_id = str( reservation.instances[0]).split(":")[1] + uci_wrapper.set_instance_id( i_index, i_id ) + s = reservation.instances[0].state + uci_wrapper.change_state( s, i_id, s ) + log.debug( "Instance of UCI '%s' started, current state: '%s'" % ( uci_wrapper.get_name(), uci_wrapper.get_state() ) ) + except boto.exception.EC2ResponseError, e: + log.error( "EC2 response error when retrieving instance information for UCI '%s': '%s'" % ( uci_wrapper.get_name(), str(e) ) ) + uci_wrapper.set_error( "EC2 response error when retrieving instance information: " + str(e), True ) + def stopUCI( self, uci_wrapper): """ Stops all of cloud instances associated with given UCI. @@ -355,8 +379,8 @@ def update( self ): """ - Runs a global status update on all instances that are in 'running', 'pending', "creating", or 'shutting-down' state. - Also, runs update on all storage volumes that are in "in-use", "creating", or 'None' state. + Runs a global status update on all instances that are in 'running', 'pending', or 'shutting-down' state. + Also, runs update on all storage volumes that are in 'in-use', 'creating', or 'None' state. Reason behind this method is to sync state of local DB and real-world resources """ log.debug( "Running general status update for EC2 UCIs..." ) @@ -385,14 +409,28 @@ a_key = uci.credentials.access_key s_key = uci.credentials.secret_key # Get connection - region = RegionInfo( None, uci.credentials.provider.region_name, uci.credentials.provider.region_endpoint ) - conn = EC2Connection( aws_access_key_id=a_key, - aws_secret_access_key=s_key, - is_secure=uci.credentials.provider.is_secure, - region=region, - path=uci.credentials.provider.path ) + try: + region = RegionInfo( None, uci.credentials.provider.region_name, uci.credentials.provider.region_endpoint ) + conn = EC2Connection( aws_access_key_id=a_key, + aws_secret_access_key=s_key, + is_secure=uci.credentials.provider.is_secure, + region=region, + path=uci.credentials.provider.path ) + except boto.exception.EC2ResponseError, e: + log.error( "Establishing connection with cloud failed: %s" % str(e) ) + uci.error( "Establishing connection with cloud failed: " + str(e) ) + uci.state( uci_states.ERROR ) + return None + # Get reservations handle for given instance - rl= conn.get_all_instances( [inst.instance_id] ) + try: + rl= conn.get_all_instances( [inst.instance_id] ) + except boto.exception.EC2ResponseError, e: + log.error( "Retrieving instance(s) from cloud for UCI '%s' failed: " % ( uci.name, str(e) ) ) + uci.error( "Retrieving instance(s) from cloud failed: " + str(e) ) + uci.state( uci_states.ERROR ) + return None + # Because EPC deletes references to reservations after a short while after instances have terminated, getting an empty list as a response to a query # typically means the instance has successfully shut down but the check was not performed in short enough amount of time. Until alternative solution # is found, below code sets state of given UCI to 'error' to indicate to the user something out of ordinary happened. @@ -409,27 +447,33 @@ # Update instance status in local DB with info from cloud provider for r in rl: for i, cInst in enumerate( r.instances ): - s = cInst.update() - log.debug( "Checking state of cloud instance '%s' associated with UCI '%s' and reservation '%s'. State='%s'" % ( cInst, uci.name, r, s ) ) - if s != inst.state: - inst.state = s - inst.flush() - # After instance has shut down, ensure UCI is marked as 'available' - if s == instance_states.TERMINATED and uci.state != uci_states.ERROR: - uci.state = uci_states.AVAILABLE - uci.launch_time = None - uci.flush() - # Making sure state of UCI is updated. Once multiple instances become associated with single UCI, this will need to be changed. - if s != uci.state and s != instance_states.TERMINATED: - uci.state = s - uci.flush() - if cInst.public_dns_name != inst.public_dns: - inst.public_dns = cInst.public_dns_name - inst.flush() - if cInst.private_dns_name != inst.private_dns: - inst.private_dns = cInst.private_dns_name - inst.flush() - + try: + s = cInst.update() + log.debug( "Checking state of cloud instance '%s' associated with UCI '%s' and reservation '%s'. State='%s'" % ( cInst, uci.name, r, s ) ) + if s != inst.state: + inst.state = s + inst.flush() + # After instance has shut down, ensure UCI is marked as 'available' + if s == instance_states.TERMINATED and uci.state != uci_states.ERROR: + uci.state = uci_states.AVAILABLE + uci.launch_time = None + uci.flush() + # Making sure state of UCI is updated. Once multiple instances become associated with single UCI, this will need to be changed. + if s != uci.state and s != instance_states.TERMINATED: + uci.state = s + uci.flush() + if cInst.public_dns_name != inst.public_dns: + inst.public_dns = cInst.public_dns_name + inst.flush() + if cInst.private_dns_name != inst.private_dns: + inst.private_dns = cInst.private_dns_name + inst.flush() + except boto.exception.EC2ResponseError, e: + log.error( "Updating status of instance(s) from cloud for UCI '%s' failed: " % ( uci.name, str(e) ) ) + uci.error( "Updating instance status from cloud failed: " + str(e) ) + uci.state( uci_states.ERROR ) + return None + def updateStore( self, store ): # Get credentials associated wit this store uci_id = store.uci_id @@ -438,35 +482,55 @@ a_key = uci.credentials.access_key s_key = uci.credentials.secret_key # Get connection - region = RegionInfo( None, uci.credentials.provider.region_name, uci.credentials.provider.region_endpoint ) - conn = EC2Connection( aws_access_key_id=a_key, + try: + region = RegionInfo( None, uci.credentials.provider.region_name, uci.credentials.provider.region_endpoint ) + conn = EC2Connection( aws_access_key_id=a_key, aws_secret_access_key=s_key, is_secure=uci.credentials.provider.is_secure, region=region, path=uci.credentials.provider.path ) + except boto.exception.EC2ResponseError, e: + log.error( "Establishing connection with cloud failed: %s" % str(e) ) + uci.error( "Establishing connection with cloud failed: " + str(e) ) + uci.state( uci_states.ERROR ) + return None + # Get reservations handle for given store - vl = conn.get_all_volumes( [store.volume_id] ) + try: + vl = conn.get_all_volumes( [store.volume_id] ) # log.debug( "Store '%s' vl: '%s'" % ( store.volume_id, vl ) ) + except boto.exception.EC2ResponseError, e: + log.error( "Retrieving volume(s) from cloud for UCI '%s' failed: " % ( uci.name, str(e) ) ) + uci.error( "Retrieving volume(s) from cloud failed: " + str(e) ) + uci.state( uci_states.ERROR ) + return None + # Update store status in local DB with info from cloud provider - if store.status != vl[0].status: - # In case something failed during creation of UCI but actual storage volume was created and yet - # UCI state remained as 'new', try to remedy this by updating UCI state here - if ( store.status == None ) and ( store.volume_id != None ): - uci.state = vl[0].status - uci.flush() - - store.status = vl[0].status - store.flush() - if store.i_id != vl[0].instance_id: - store.i_id = vl[0].instance_id - store.flush() - if store.attach_time != vl[0].attach_time: - store.attach_time = vl[0].attach_time - store.flush() - if store.device != vl[0].device: - store.device = vl[0].device - store.flush() - + try: + if store.status != vl[0].status: + # In case something failed during creation of UCI but actual storage volume was created and yet + # UCI state remained as 'new', try to remedy this by updating UCI state here + if ( store.status == None ) and ( store.volume_id != None ): + uci.state = vl[0].status + uci.flush() + + store.status = vl[0].status + store.flush() + if store.i_id != vl[0].instance_id: + store.i_id = vl[0].instance_id + store.flush() + if store.attach_time != vl[0].attach_time: + store.attach_time = vl[0].attach_time + store.flush() + if store.device != vl[0].device: + store.device = vl[0].device + store.flush() + except boto.exception.EC2ResponseError, e: + log.error( "Updating status of volume(s) from cloud for UCI '%s' failed: " % ( uci.name, str(e) ) ) + uci.error( "Updating volume status from cloud failed: " + str(e) ) + uci.state( uci_states.ERROR ) + return None + # def updateUCI( self, uci ): # """ # Runs a global status update on all storage volumes and all instances that are diff -r be631ed97541 -r e19eef93584f lib/galaxy/cloud/providers/eucalyptus.py --- a/lib/galaxy/cloud/providers/eucalyptus.py Thu Oct 29 11:46:43 2009 -0400 +++ b/lib/galaxy/cloud/providers/eucalyptus.py Thu Oct 29 17:40:31 2009 -0400 @@ -7,7 +7,7 @@ from galaxy.datatypes.data import nice_size from galaxy.util.bunch import Bunch from Queue import Queue -from sqlalchemy import or_ +from sqlalchemy import or_, and_ import galaxy.eggs galaxy.eggs.require("boto") @@ -69,9 +69,8 @@ """Run the next job, waiting until one is available if necessary""" cnt = 0 while 1: - + #log.debug( '[%d] run_next->queue.qsize(): %s' % ( cnt, self.queue.qsize() ) ) uci_wrapper = self.queue.get() - log.debug( '[%d] uci type: %s' % ( cnt, uci_wrapper.get_name() ) ) uci_state = uci_wrapper.get_state() if uci_state is self.STOP_SIGNAL: return @@ -82,6 +81,7 @@ self.deleteUCI( uci_wrapper ) elif uci_state==uci_states.SUBMITTED: self.startUCI( uci_wrapper ) + #self.dummyStartUCI( uci_wrapper ) elif uci_state==uci_states.SHUTTING_DOWN: self.stopUCI( uci_wrapper ) except: @@ -94,13 +94,25 @@ """ log.debug( '##### Establishing eucalyptus cloud connection' ) provider = uci_wrapper.get_provider() - euca_region = RegionInfo( None, provider.region_name, provider.region_endpoint ) - conn = EC2Connection( aws_access_key_id=uci_wrapper.get_access_key(), - aws_secret_access_key=uci_wrapper.get_secret_key(), - is_secure=provider.is_secure, - port=provider.port, - region=euca_region, - path=provider.path ) + try: + euca_region = RegionInfo( None, provider.region_name, provider.region_endpoint ) + except Exception, e: + log.error( "Selecting region with cloud provider failed: %2" % str(e) ) + uci_wrapper.set_error( "Selecting region with cloud provider failed: " + str(e), True ) + return None + + try: + conn = EC2Connection( aws_access_key_id=uci_wrapper.get_access_key(), + aws_secret_access_key=uci_wrapper.get_secret_key(), + is_secure=provider.is_secure, + port=provider.port, + region=euca_region, + path=provider.path ) + except boto.exception.EC2ResponseError, e: + log.error( "Establishing connection with cloud failed: %s" % str(e) ) + uci_wrapper.set_error( "Establishing connection with cloud failed: " + str(e), True ) + return None + return conn def set_keypair( self, uci_wrapper, conn ): @@ -108,20 +120,22 @@ Generate keypair using user's default credentials """ log.debug( "Getting user's keypair: '%s'" % self.key_pair ) - kp = conn.get_key_pair( self.key_pair ) instances = uci_wrapper.get_instances_indexes() - try: + kp = conn.get_key_pair( self.key_pair ) for inst in instances: - log.debug("inst: '%s'" % inst ) uci_wrapper.set_key_pair( inst, kp.name ) return kp.name - except AttributeError: # No keypair under this name exists so create it - log.info( "No keypair found, creating keypair '%s'" % self.key_pair ) - kp = conn.create_key_pair( self.key_pair ) - for inst in instances: - uci_wrapper.set_key_pair( inst, kp.name, kp.material ) - + except boto.exception.EC2ResponseError, e: # No keypair under this name exists so create it + if e.code == 'InvalidKeyPair.NotFound': + log.info( "No keypair found, creating keypair '%s'" % self.key_pair ) + kp = conn.create_key_pair( self.key_pair ) + for inst in instances: + uci_wrapper.set_key_pair( inst, kp.name, kp.material ) + else: + log.error( "EC2 response error: '%s'" % e ) + uci_wrapper.set_error( "EC2 response error while creating key pair: " + str(e), True ) + return kp.name def get_mi_id( self, type ): @@ -208,10 +222,11 @@ uci = uci_wrapper.get_uci() log.debug( "Would be starting instance '%s'" % uci.name ) - uci_wrapper.change_state( uci_states.PENDING ) -# log.debug( "Sleeping a bit... (%s)" % uci.name ) -# time.sleep(20) -# log.debug( "Woke up! (%s)" % uci.name ) +# uci_wrapper.change_state( uci_states.SUBMITTED_UCI ) +# log.debug( "Set UCI state to SUBMITTED_UCI" ) + log.debug( "Sleeping a bit... (%s)" % uci.name ) + time.sleep(10) + log.debug( "Woke up! (%s)" % uci.name ) def startUCI( self, uci_wrapper ): """ @@ -219,30 +234,43 @@ """ conn = self.get_connection( uci_wrapper ) # - self.set_keypair( uci_wrapper, conn ) + if uci_wrapper.get_state() != uci_states.ERROR: + self.set_keypair( uci_wrapper, conn ) i_indexes = uci_wrapper.get_instances_indexes() # Get indexes of i_indexes associated with this UCI - for i_index in i_indexes: - mi_id = self.get_mi_id( uci_wrapper.get_type( i_index ) ) - log.debug( "mi_id: %s, uci_wrapper.get_key_pair_name( i_index ): %s" % ( mi_id, uci_wrapper.get_key_pair_name( i_index ) ) ) - uci_wrapper.set_mi( i_index, mi_id ) + if uci_wrapper.get_state() != uci_states.ERROR: + for i_index in i_indexes: + mi_id = self.get_mi_id( uci_wrapper.get_type( i_index ) ) + log.debug( "mi_id: %s, uci_wrapper.get_key_pair_name( i_index ): %s" % ( mi_id, uci_wrapper.get_key_pair_name( i_index ) ) ) + uci_wrapper.set_mi( i_index, mi_id ) - log.debug( "***** Starting UCI instance '%s'" % uci_wrapper.get_name() ) - log.debug( 'Using following command: conn.run_instances( image_id=%s, key_name=%s )' % ( mi_id, uci_wrapper.get_key_pair_name( i_index ) ) ) - reservation = conn.run_instances( image_id=mi_id, key_name=uci_wrapper.get_key_pair_name( i_index ) ) - #reservation = conn.run_instances( image_id=instance.image, key_name=instance.keypair_name, security_groups=['galaxy'], instance_type=instance.type, placement=instance.availability_zone ) - l_time = datetime.utcnow() - uci_wrapper.set_launch_time( l_time, i_index=i_index ) # format_time( reservation.i_indexes[0].launch_time ) ) - if not uci_wrapper.uci_launch_time_set(): - uci_wrapper.set_uci_launch_time( l_time ) - uci_wrapper.set_reservation_id( i_index, str( reservation ).split(":")[1] ) - # TODO: if more than a single instance will be started through single reservation, change this reference to element [0] - i_id = str( reservation.instances[0]).split(":")[1] - uci_wrapper.set_instance_id( i_index, i_id ) - s = reservation.instances[0].state - uci_wrapper.change_state( s, i_id, s ) - log.debug( "Instance of UCI '%s' started, current state: '%s'" % ( uci_wrapper.get_name(), uci_wrapper.get_state() ) ) + if uci_wrapper.get_state() != uci_states.ERROR: + log.debug( "***** Starting UCI instance '%s'" % uci_wrapper.get_name() ) + log.debug( 'Using following command: conn.run_instances( image_id=%s, key_name=%s )' % ( mi_id, uci_wrapper.get_key_pair_name( i_index ) ) ) + try: + reservation = conn.run_instances( image_id=mi_id, key_name=uci_wrapper.get_key_pair_name( i_index ) ) + #reservation = conn.run_instances( image_id=instance.image, key_name=instance.keypair_name, security_groups=['galaxy'], instance_type=instance.type, placement=instance.availability_zone ) + except boto.exception.EC2ResponseError, e: + log.error( "EC2 response error when starting UCI '%s': '%s'" % ( uci_wrapper.get_name(), str(e) ) ) + uci_wrapper.set_error( "EC2 response error when starting: " + str(e), True ) + + l_time = datetime.utcnow() + uci_wrapper.set_launch_time( l_time, i_index=i_index ) # format_time( reservation.i_indexes[0].launch_time ) ) + if not uci_wrapper.uci_launch_time_set(): + uci_wrapper.set_uci_launch_time( l_time ) + try: + uci_wrapper.set_reservation_id( i_index, str( reservation ).split(":")[1] ) + # TODO: if more than a single instance will be started through single reservation, change this reference to element [0] + i_id = str( reservation.instances[0]).split(":")[1] + uci_wrapper.set_instance_id( i_index, i_id ) + s = reservation.instances[0].state + uci_wrapper.change_state( s, i_id, s ) + log.debug( "Instance of UCI '%s' started, current state: '%s'" % ( uci_wrapper.get_name(), uci_wrapper.get_state() ) ) + except boto.exception.EC2ResponseError, e: + log.error( "EC2 response error when retrieving instance information for UCI '%s': '%s'" % ( uci_wrapper.get_name(), str(e) ) ) + uci_wrapper.set_error( "EC2 response error when retrieving instance information: " + str(e), True ) + def stopUCI( self, uci_wrapper): """ @@ -316,11 +344,12 @@ def update( self ): """ - Runs a global status update on all instances that are in 'running', 'pending', "creating", or 'shutting-down' state. - Also, runs update on all storage volumes that are in "in-use", "creating", or 'None' state. + Runs a global status update on all instances that are in 'running', 'pending', or 'shutting-down' state. + Also, runs update on all storage volumes that are in 'in-use', 'creating', or 'None' state. Reason behind this method is to sync state of local DB and real-world resources """ log.debug( "Running general status update for EPC UCIs..." ) + # Update instances instances = model.CloudInstance.filter( or_( model.CloudInstance.c.state==instance_states.RUNNING, model.CloudInstance.c.state==instance_states.PENDING, model.CloudInstance.c.state==instance_states.SHUTTING_DOWN ) ).all() @@ -328,7 +357,8 @@ if self.type == inst.uci.credentials.provider.type: log.debug( "[%s] Running general status update on instance '%s'" % ( inst.uci.credentials.provider.type, inst.instance_id ) ) self.updateInstance( inst ) - + + # Update storage volume(s) stores = model.CloudStore.filter( or_( model.CloudStore.c.status==store_states.IN_USE, model.CloudStore.c.status==store_states.CREATING, model.CloudStore.c.status==None ) ).all() @@ -337,6 +367,20 @@ log.debug( "[%s] Running general status update on store '%s'" % ( store.uci.credentials.provider.type, store.volume_id ) ) self.updateStore( store ) + # Attempt at updating any zombie UCIs (i.e., instances that have been in SUBMITTED state for longer than expected - see below for exact time) + zombies = model.UCI.filter_by( state=uci_states.SUBMITTED ).all() + for zombie in zombies: + z_instances = model.CloudInstance.filter_by( uci_id=zombie.id) \ + .filter( or_( model.CloudInstance.c.state!=instance_states.TERMINATED, + model.CloudInstance.c.state == None ) ) \ + .all() + for z_inst in z_instances: + if self.type == z_inst.uci.credentials.provider.type: + log.debug( "z_inst.id: '%s', state: '%s'" % ( z_inst.id, z_inst.state ) ) +# td = datetime.utcnow() - zombie.update_time +# if td.seconds > 180: # if instance has been in SUBMITTED state for more than 3 minutes +# log.debug( "[%s] Running zombie repair update on instance '%s'" % ( inst.uci.credentials.provider.type, inst.id ) ) + def updateInstance( self, inst ): # Get credentials associated wit this instance @@ -346,15 +390,29 @@ a_key = uci.credentials.access_key s_key = uci.credentials.secret_key # Get connection - euca_region = RegionInfo( None, uci.credentials.provider.region_name, uci.credentials.provider.region_endpoint ) - conn = EC2Connection( aws_access_key_id=a_key, - aws_secret_access_key=s_key, - is_secure=uci.credentials.provider.is_secure, - port=uci.credentials.provider.port, - region=euca_region, - path=uci.credentials.provider.path ) + try: + euca_region = RegionInfo( None, uci.credentials.provider.region_name, uci.credentials.provider.region_endpoint ) + conn = EC2Connection( aws_access_key_id=a_key, + aws_secret_access_key=s_key, + is_secure=uci.credentials.provider.is_secure, + port=uci.credentials.provider.port, + region=euca_region, + path=uci.credentials.provider.path ) + except boto.exception.EC2ResponseError, e: + log.error( "Establishing connection with cloud failed: %s" % str(e) ) + uci.error( "Establishing connection with cloud failed: " + str(e) ) + uci.state( uci_states.ERROR ) + return None + # Get reservations handle for given instance - rl= conn.get_all_instances( [inst.instance_id] ) + try: + rl= conn.get_all_instances( [inst.instance_id] ) + except boto.exception.EC2ResponseError, e: + log.error( "Retrieving instance(s) from cloud for UCI '%s' failed: " % ( uci.name, str(e) ) ) + uci.error( "Retrieving instance(s) from cloud failed: " + str(e) ) + uci.state( uci_states.ERROR ) + return None + # Because EPC deletes references to reservations after a short while after instances have terminated, getting an empty list as a response to a query # typically means the instance has successfully shut down but the check was not performed in short enough amount of time. Until alternative solution # is found, below code sets state of given UCI to 'error' to indicate to the user something out of ordinary happened. @@ -371,26 +429,32 @@ # Update instance status in local DB with info from cloud provider for r in rl: for i, cInst in enumerate( r.instances ): - s = cInst.update() - log.debug( "Checking state of cloud instance '%s' associated with reservation '%s'. State='%s'" % ( cInst, r, s ) ) - if s != inst.state: - inst.state = s + try: + s = cInst.update() + log.debug( "Checking state of cloud instance '%s' associated with reservation '%s'. State='%s'" % ( cInst, r, s ) ) + if s != inst.state: + inst.state = s + inst.flush() + # After instance has shut down, ensure UCI is marked as 'available' + if s == instance_states.TERMINATED and uci.state != uci_states.ERROR: + uci.state = uci_states.AVAILABLE + uci.launch_time = None + uci.flush() + # Making sure state of UCI is updated. Once multiple instances become associated with single UCI, this will need to be changed. + if s != uci.state and s != instance_states.TERMINATED: + uci.state = s + uci.flush() + if cInst.public_dns_name != inst.public_dns: + inst.public_dns = cInst.public_dns_name + inst.flush() + if cInst.private_dns_name != inst.private_dns: + inst.private_dns = cInst.private_dns_name inst.flush() - # After instance has shut down, ensure UCI is marked as 'available' - if s == instance_states.TERMINATED and uci.state != uci_states.ERROR: - uci.state = uci_states.AVAILABLE - uci.launch_time = None - uci.flush() - # Making sure state of UCI is updated. Once multiple instances become associated with single UCI, this will need to be changed. - if s != uci.state and s != instance_states.TERMINATED: - uci.state = s - uci.flush() - if cInst.public_dns_name != inst.public_dns: - inst.public_dns = cInst.public_dns_name - inst.flush() - if cInst.private_dns_name != inst.private_dns: - inst.private_dns = cInst.private_dns_name - inst.flush() + except boto.exception.EC2ResponseError, e: + log.error( "Updating status of instance(s) from cloud for UCI '%s' failed: " % ( uci.name, str(e) ) ) + uci.error( "Updating volume status from cloud failed: " + str(e) ) + uci.state( uci_states.ERROR ) + return None def updateStore( self, store ): # Get credentials associated wit this store @@ -400,34 +464,53 @@ a_key = uci.credentials.access_key s_key = uci.credentials.secret_key # Get connection - euca_region = RegionInfo( None, uci.credentials.provider.region_name, uci.credentials.provider.region_endpoint ) - conn = EC2Connection( aws_access_key_id=a_key, - aws_secret_access_key=s_key, - is_secure=uci.credentials.provider.is_secure, - port=uci.credentials.provider.port, - region=euca_region, - path=uci.credentials.provider.path )# Get reservations handle for given store - vl = conn.get_all_volumes( [store.volume_id] ) + try: + euca_region = RegionInfo( None, uci.credentials.provider.region_name, uci.credentials.provider.region_endpoint ) + conn = EC2Connection( aws_access_key_id=a_key, + aws_secret_access_key=s_key, + is_secure=uci.credentials.provider.is_secure, + port=uci.credentials.provider.port, + region=euca_region, + path=uci.credentials.provider.path ) + except boto.exception.EC2ResponseError, e: + log.error( "Establishing connection with cloud failed: %s" % str(e) ) + uci.error( "Establishing connection with cloud failed: " + str(e) ) + uci.state( uci_states.ERROR ) + return None + + try: + vl = conn.get_all_volumes( [store.volume_id] ) + except boto.exception.EC2ResponseError, e: + log.error( "Retrieving volume(s) from cloud for UCI '%s' failed: " % ( uci.name, str(e) ) ) + uci.error( "Retrieving volume(s) from cloud failed: " + str(e) ) + uci.state( uci_states.ERROR ) + return None + # Update store status in local DB with info from cloud provider - if store.status != vl[0].status: - # In case something failed during creation of UCI but actual storage volume was created and yet - # UCI state remained as 'new', try to remedy this by updating UCI state here - if ( store.status == None ) and ( store.volume_id != None ): - uci.state = vl[0].status - uci.flush() - - store.status = vl[0].status - store.flush() - if store.i_id != vl[0].instance_id: - store.i_id = vl[0].instance_id - store.flush() - if store.attach_time != vl[0].attach_time: - store.attach_time = vl[0].attach_time - store.flush() - if store.device != vl[0].device: - store.device = vl[0].device - store.flush() - + try: + if store.status != vl[0].status: + # In case something failed during creation of UCI but actual storage volume was created and yet + # UCI state remained as 'new', try to remedy this by updating UCI state here + if ( store.status == None ) and ( store.volume_id != None ): + uci.state = vl[0].status + uci.flush() + + store.status = vl[0].status + store.flush() + if store.i_id != vl[0].instance_id: + store.i_id = vl[0].instance_id + store.flush() + if store.attach_time != vl[0].attach_time: + store.attach_time = vl[0].attach_time + store.flush() + if store.device != vl[0].device: + store.device = vl[0].device + store.flush() + except boto.exception.EC2ResponseError, e: + log.error( "Updating status of volume(s) from cloud for UCI '%s' failed: " % ( uci.name, str(e) ) ) + uci.error( "Updating volume status from cloud failed: " + str(e) ) + uci.state( uci_states.ERROR ) + return None # def updateUCI( self, uci ): # """ # Runs a global status update on all storage volumes and all instances that are diff -r be631ed97541 -r e19eef93584f lib/galaxy/web/controllers/cloud.py --- a/lib/galaxy/web/controllers/cloud.py Thu Oct 29 11:46:43 2009 -0400 +++ b/lib/galaxy/web/controllers/cloud.py Thu Oct 29 17:40:31 2009 -0400 @@ -104,9 +104,9 @@ model.UCI.c.state==uci_states.SUBMITTED_UCI ) ) \ .all() if pendingInstances: - trans.set_message( "Galaxy instance started. NOTE: Please wait about 3-5 minutes for the instance to " - "start up and then refresh this page. A button to connect to the instance will then appear alongside " - "instance description." ) + trans.set_message( "Galaxy instance started. Note that it will take several minutes for the instance to start " + "(typically, 3-5 minutes). Once the instance is running and Galaxy is available, " + "a button to connect to the instance will then appear alongside instance description." ) # log.debug( "provider.is_secure: '%s'" % trans.sa_session.query( model.CloudProvider).filter_by(id=1).first().is_secure ) # trans.sa_session.query( model.CloudProvider).filter_by(id=1).first().is_secure=False @@ -151,17 +151,7 @@ stores = get_stores( trans, uci ) # Ensure instance is not already running (or related state) and store relevant data # into DB to initiate instance startup by cloud manager - if ( len(stores) is not 0 ) and \ - ( uci.state != uci_states.SUBMITTED ) and \ - ( uci.state != uci_states.SUBMITTED_UCI ) and \ - ( uci.state != uci_states.PENDING ) and \ - ( uci.state != uci_states.DELETING ) and \ - ( uci.state != uci_states.DELETING_UCI ) and \ - ( uci.state != uci_states.DELETED ) and \ - ( uci.state != uci_states.RUNNING ) and \ - ( uci.state != uci_states.NEW_UCI ) and \ - ( uci.state != uci_states.NEW ) and \ - ( uci.state != uci_states.ERROR ): + if ( len(stores) is not 0 ) and ( uci.state == uci_states.AVAILABLE ): instance = model.CloudInstance() instance.user = user instance.image = mi @@ -175,7 +165,7 @@ session.save_or_update( uci ) session.flush() # Log - trans.log_event ("User initiated starting of cloud instance '%s'." % uci.name ) + trans.log_event ("User initiated starting of UCI '%s'." % uci.name ) trans.set_message( "Galaxy instance started. NOTE: Please wait about 3-5 minutes for the instance to " "start up and then refresh this page. A button to connect to the instance will then appear alongside " "instance description." ) @@ -268,14 +258,20 @@ storedCreds = trans.sa_session.query( model.CloudUserCredentials ).filter_by( user=user ).all() if len( storedCreds ) == 0: return trans.show_error_message( "You must register credentials before configuring a Galaxy instance." ) + # TODO: This should be filled automatically but ties to implementation for diff provider is a problem... # Create dict mapping of cloud providers to zones available by those providers providersToZones = {} for storedCred in storedCreds: - if storedCred.provider.type == 'ec2': + if storedCred.provider.region_name == 'us-east-1': ec2_zones = ['us-east-1a', 'us-east-1b', 'us-east-1c', 'us-east-1d'] providersToZones[storedCred.name] = ec2_zones + elif storedCred.provider.region_name == 'eu-west-1': + ec2_zones = ['eu-west-1a', 'eu-west-1b'] + providersToZones[storedCred.name] = ec2_zones elif storedCred.provider.type == 'eucalyptus': providersToZones[storedCred.name] = ['epc'] + else: + providersToZones[storedCred.name] = ['Unknown provider zone'] if instanceName: # Create new user configured instance diff -r be631ed97541 -r e19eef93584f templates/cloud/add_provider.mako --- a/templates/cloud/add_provider.mako Thu Oct 29 11:46:43 2009 -0400 +++ b/templates/cloud/add_provider.mako Thu Oct 29 17:40:31 2009 -0400 @@ -117,9 +117,6 @@ <div id="region_selection" class="form-row-input"> <input type="text" name="region_name" id="region_name" value="${region_name}" size="40"> </div> - %if error.has_key('name_error'): - <div class="form-row-error-message">${error['name_error']}</div> - %endif <div style="clear: both"></div> </div> diff -r be631ed97541 -r e19eef93584f templates/cloud/configure_cloud.mako --- a/templates/cloud/configure_cloud.mako Thu Oct 29 11:46:43 2009 -0400 +++ b/templates/cloud/configure_cloud.mako Thu Oct 29 17:40:31 2009 -0400 @@ -33,6 +33,11 @@ if ( old_state=='pending' && new_state=='running' ) { location.reload(true); } + else if ( ( old_state=='running' && new_state=='error' ) || ( old_state=='pending' && new_state=='error' ) || \ + ( old_state=='submitted' && new_state=='error' ) || ( old_state=='submittedUCI' && new_state=='error' ) || \ + ( old_state=='shutting-down' && new_state=='error' ) ) { + location.reload(true); + } else if ( old_state=='shutting-down' && new_state=='available' ) { location.reload(true); } @@ -244,14 +249,14 @@ %if state =='error': <div id="short"> <a onclick="document.getElementById('full').style.display = 'block'; - document.getElementById('short').style.display = 'none'; return 0" + document.getElementById('short').style.display = 'none'; return 0" href="javascript:void(0)"> error </a> </div> <div id="full" style="DISPLAY: none"> <a onclick="document.getElementById('short').style.display = 'block'; - document.getElementById('full').style.display = 'none'; return 0;" + document.getElementById('full').style.display = 'none'; return 0;" href="javascript:void(0)"> error:</a><br /> ${str(prevInstance.error)} diff -r be631ed97541 -r e19eef93584f templates/cloud/configure_uci.mako --- a/templates/cloud/configure_uci.mako Thu Oct 29 11:46:43 2009 -0400 +++ b/templates/cloud/configure_uci.mako Thu Oct 29 17:40:31 2009 -0400 @@ -81,7 +81,7 @@ cls += " form-row-error" %> <div class="${cls}"> - <label>Permanent storage size (1-1000GB)<br/>NOTE: you will be able to add more storage later:</label> + <label>Permanent storage size (1-1000GB):<br/>(Note: you will be able to add more storage later)</label> <div class="form-row-input"> <input type="text" name="volSize" value="${volSize}" size="40"> </div> @@ -97,7 +97,7 @@ cls += " form-row-error" %> <div class="${cls}"> - <label>Zone to create storage in</label> + <label>Zone to create storage in:</label> <div class="form-row-input"> <select id="zones" name="zone" style="width:40em"> </select> diff -r be631ed97541 -r e19eef93584f templates/cloud/viewInstance.mako --- a/templates/cloud/viewInstance.mako Thu Oct 29 11:46:43 2009 -0400 +++ b/templates/cloud/viewInstance.mako Thu Oct 29 17:40:31 2009 -0400 @@ -83,12 +83,15 @@ </tr> <tr> <td> Public DNS:</td> - <td> ${liveInstance.public_dns} </td> + <% + lnk="http://"+str(liveInstance.public_dns) + %> + <td> <a href="${lnk}" target="_blank">${liveInstance.public_dns}</a></td> </tr> %if liveInstance.private_dns != None: <tr> <td> Private DNS:</td> - <td> ${liveInstance.private_dns} </td> + <td> ${liveInstance.private_dns}</td> </tr> %endif %if liveInstance.availability_zone != None:

1 0

[hg] galaxy 3077: Cleaned up code a bit (i.e., deleted temporary...
by Greg Von Kuster 23 Nov '09

23 Nov '09

details: http://www.bx.psu.edu/hg/galaxy/rev/be631ed97541 changeset: 3077:be631ed97541 user: Enis Afgan <afgane(a)gmail.com> date: Thu Oct 29 11:46:43 2009 -0400 description: Cleaned up code a bit (i.e., deleted temporary code/comments). diffstat: lib/galaxy/cloud/__init__.py | 701 +------------------------------------ lib/galaxy/cloud/providers/ec2.py | 93 +---- lib/galaxy/cloud/providers/eucalyptus.py | 100 +---- lib/galaxy/web/controllers/cloud.py | 235 ------------ 4 files changed, 23 insertions(+), 1106 deletions(-) diffs (1457 lines): diff -r 022ac0f64679 -r be631ed97541 lib/galaxy/cloud/__init__.py --- a/lib/galaxy/cloud/__init__.py Wed Oct 28 16:35:36 2009 -0400 +++ b/lib/galaxy/cloud/__init__.py Thu Oct 29 11:46:43 2009 -0400 @@ -47,40 +47,11 @@ self.provider = CloudProvider( app ) # Monitor for updating status of cloud instances self.cloud_monitor = CloudMonitor( self.app, self.provider ) -# self.job_stop_queue = JobStopQueue( app, self.dispatcher ) else: self.job_queue = self.job_stop_queue = NoopCloudMonitor() def shutdown( self ): self.cloud_monitor.shutdown() -# self.job_stop_queue.shutdown() - -# def createUCI( self, user, name, storage_size, zone=None): -# """ -# Createse User Configured Instance (UCI). Essentially, creates storage volume. -# """ -# self.provider.createUCI( user, name, storage_size, zone ) -# -# def deleteUCI( self, name ): -# """ -# Deletes UCI. NOTE that this implies deletion of any and all data associated -# with this UCI from the cloud. All data will be deleted. -# """ -# -# def addStorageToUCI( self, name ): -# """ Adds more storage to specified UCI """ -# -# def startUCI( self, name, type ): -# """ -# Starts an instance of named UCI on the cloud. This implies, mounting of -# storage and starting Galaxy instance. -# """ -# -# def stopUCI( self, name ): -# """ -# Stops cloud instance associated with named UCI. This also implies -# stopping of Galaxy and unmounting of the file system. -# """ class Sleeper( object ): """ @@ -110,9 +81,7 @@ # Keep track of the pid that started the cloud manager, only it # has valid threads self.parent_pid = os.getpid() - # Contains new jobs. Note this is not used if track_jobs_in_database is True -# self.queue = Queue() - + # Contains requests that are waiting (only use from monitor thread) self.waiting = [] @@ -123,33 +92,6 @@ self.monitor_thread = threading.Thread( target=self.__monitor ) self.monitor_thread.start() log.info( "Cloud manager started" ) -# if app.config.get_bool( 'enable_job_recovery', True ): -# self.__check_jobs_at_startup() - - def __check_jobs_at_startup( self ): - """ - Checks all jobs that are in the 'new', 'queued' or 'running' state in - the database and requeues or cleans up as necessary. Only run as the - job manager starts. - """ - model = self.app.model - for job in model.Job.filter( model.Job.c.state==model.Job.states.NEW ).all(): - if job.tool_id not in self.app.toolbox.tools_by_id: - log.warning( "Tool '%s' removed from tool config, unable to recover job: %s" % ( job.tool_id, job.id ) ) - JobWrapper( job, None, self ).fail( 'This tool was disabled before the job completed. Please contact your Galaxy administrator, or' ) - else: - log.debug( "no runner: %s is still in new state, adding to the jobs queue" %job.id ) - self.queue.put( ( job.id, job.tool_id ) ) - for job in model.Job.filter( (model.Job.c.state == model.Job.states.RUNNING) | (model.Job.c.state == model.Job.states.QUEUED) ).all(): - if job.tool_id not in self.app.toolbox.tools_by_id: - log.warning( "Tool '%s' removed from tool config, unable to recover job: %s" % ( job.tool_id, job.id ) ) - JobWrapper( job, None, self ).fail( 'This tool was disabled before the job completed. Please contact your Galaxy administrator, or' ) - elif job.job_runner_name is None: - log.debug( "no runner: %s is still in queued state, adding to the jobs queue" %job.id ) - self.queue.put( ( job.id, job.tool_id ) ) - else: - job_wrapper = JobWrapper( job, self.app.toolbox.tools_by_id[ job.tool_id ], self ) - self.dispatcher.recover( job, job_wrapper ) def __monitor( self ): """ @@ -164,7 +106,7 @@ try: # log.debug( "Calling monitor_step" ) self.__monitor_step() - if cnt%30 == 0: # Run global update every 30 seconds (1 minute) + if cnt%30 == 0: # Run global update every 30 iterations (1 minute) self.provider.update() cnt = 0 except: @@ -186,166 +128,24 @@ """ # Get an orm (object relational mapping) session session = mapping.Session() - # Pull all new jobs from the queue at once new_requests = [] -# new_instances = [] -# new_UCIs = [] -# stop_UCIs = [] -# delete_UCIs = [] - -# for r in session.query( model.cloud_instance ).filter( model.cloud_instance.s.state == model.cloud_instance.states.NEW ).all(): -# new_instances - + for r in session.query( model.UCI ) \ - .filter( or_( model.UCI.c.state==uci_states.NEW_UCI, #"newUCI", - model.UCI.c.state==uci_states.SUBMITTED_UCI, #"submittedUCI", - model.UCI.c.state==uci_states.SHUTTING_DOWN_UCI, #"shutting-downUCI", + .filter( or_( model.UCI.c.state==uci_states.NEW_UCI, + model.UCI.c.state==uci_states.SUBMITTED_UCI, + model.UCI.c.state==uci_states.SHUTTING_DOWN_UCI, model.UCI.c.state==uci_states.DELETING_UCI ) ) \ .all(): uci_wrapper = UCIwrapper( r ) new_requests.append( uci_wrapper ) -# log.debug( 'new_requests: %s' % new_requests ) + for uci_wrapper in new_requests: session.clear() -# log.debug( 'r.name: %s, state: %s' % ( r.name, r.state ) ) -# session.save_or_update( r ) -# session.flush() self.provider.put( uci_wrapper ) # Done with the session mapping.Session.remove() - -# for r in session.query( model.UCI ).filter( model.UCI.c.state == "submitted" ).all(): -# new_instances.append( r ) -# for r in new_instances: -# self.provider.startUCI( r ) -# -# for r in session.query( model.UCI ).filter( model.UCI.c.state == "shutting-down" ).all(): -# stop_UCIs.append( r ) -# for r in stop_UCIs: -# self.provider.stopUCI( r ) -# -# for r in session.query( model.UCI ).filter( model.UCI.c.state == "deleting" ).all(): -# delete_UCIs.append( r ) -# for r in delete_UCIs: -# self.provider.deleteUCI( r ) - - - -# if self.track_jobs_in_database: -# for j in session.query( model.Job ).options( lazyload( "external_output_metadata" ), lazyload( "parameters" ) ).filter( model.Job.c.state == model.Job.states.NEW ).all(): -# job = JobWrapper( j, self.app.toolbox.tools_by_id[ j.tool_id ], self ) -# new_jobs.append( job ) -# else: -# try: -# while 1: -# message = self.queue.get_nowait() -# if message is self.STOP_SIGNAL: -# return -# # Unpack the message -# job_id, tool_id = message -# # Create a job wrapper from it -# job_entity = session.query( model.Job ).get( job_id ) -# job = JobWrapper( job_entity, self.app.toolbox.tools_by_id[ tool_id ], self ) -# # Append to watch queue -# new_jobs.append( job ) -# except Empty: -# pass -# # Iterate over new and waiting jobs and look for any that are -# # ready to run -# new_waiting = [] -# for job in ( new_jobs + self.waiting ): -# try: -# # Clear the session for each job so we get fresh states for -# # job and all datasets -# session.clear() -# # Get the real job entity corresponding to the wrapper (if we -# # are tracking in the database this is probably cached in -# # the session from the origianl query above) -# job_entity = session.query( model.Job ).get( job.job_id ) -# # Check the job's dependencies, requeue if they're not done -# job_state = self.__check_if_ready_to_run( job, job_entity ) -# if job_state == JOB_WAIT: -# if not self.track_jobs_in_database: -# new_waiting.append( job ) -# elif job_state == JOB_ERROR: -# log.info( "job %d ended with an error" % job.job_id ) -# elif job_state == JOB_INPUT_ERROR: -# log.info( "job %d unable to run: one or more inputs in error state" % job.job_id ) -# elif job_state == JOB_INPUT_DELETED: -# log.info( "job %d unable to run: one or more inputs deleted" % job.job_id ) -# elif job_state == JOB_READY: -# # If special queuing is enabled, put the ready jobs in the special queue -# if self.use_policy : -# self.squeue.put( job ) -# log.debug( "job %d put in policy queue" % job.job_id ) -# else: # or dispatch the job directly -# self.dispatcher.put( job ) -# log.debug( "job %d dispatched" % job.job_id) -# elif job_state == JOB_DELETED: -# msg = "job %d deleted by user while still queued" % job.job_id -# job.info = msg -# log.debug( msg ) -# elif job_state == JOB_ADMIN_DELETED: -# job.fail( job_entity.info ) -# log.info( "job %d deleted by admin while still queued" % job.job_id ) -# else: -# msg = "unknown job state '%s' for job %d" % ( job_state, job.job_id ) -# job.info = msg -# log.error( msg ) -# except Exception, e: -# job.info = "failure running job %d: %s" % ( job.job_id, str( e ) ) -# log.exception( "failure running job %d" % job.job_id ) -# # Update the waiting list -# self.waiting = new_waiting -# # If special (e.g. fair) scheduling is enabled, dispatch all jobs -# # currently in the special queue -# if self.use_policy : -# while 1: -# try: -# sjob = self.squeue.get() -# self.dispatcher.put( sjob ) -# log.debug( "job %d dispatched" % sjob.job_id ) -# except Empty: -# # squeue is empty, so stop dispatching -# break -# except Exception, e: # if something else breaks while dispatching -# job.fail( "failure running job %d: %s" % ( sjob.job_id, str( e ) ) ) -# log.exception( "failure running job %d" % sjob.job_id ) -# # Done with the session -# mapping.Session.remove() - - def __check_if_ready_to_run( self, job_wrapper, job ): - """ - Check if a job is ready to run by verifying that each of its input - datasets is ready (specifically in the OK state). If any input dataset - has an error, fail the job and return JOB_INPUT_ERROR. If any input - dataset is deleted, fail the job and return JOB_INPUT_DELETED. If all - input datasets are in OK state, return JOB_READY indicating that the - job can be dispatched. Otherwise, return JOB_WAIT indicating that input - datasets are still being prepared. - """ - if job.state == model.Job.states.DELETED: - return JOB_DELETED - elif job.state == model.Job.states.ERROR: - return JOB_ADMIN_DELETED - for dataset_assoc in job.input_datasets: - idata = dataset_assoc.dataset - if not idata: - continue - # don't run jobs for which the input dataset was deleted - if idata.deleted: - job_wrapper.fail( "input data %d (file: %s) was deleted before the job started" % ( idata.hid, idata.file_name ) ) - return JOB_INPUT_DELETED - # an error in the input data causes us to bail immediately - elif idata.state == idata.states.ERROR: - job_wrapper.fail( "input data %d is in error state" % ( idata.hid ) ) - return JOB_INPUT_ERROR - elif idata.state != idata.states.OK and not ( idata.state == idata.states.SETTING_METADATA and job.tool_id is not None and job.tool_id == self.app.datatypes_registry.set_external_metadata_tool.id ): - # need to requeue - return JOB_WAIT - return JOB_READY - + def put( self, job_id, tool ): """Add a job to the queue (by job identifier)""" if not self.track_jobs_in_database: @@ -355,13 +155,11 @@ def shutdown( self ): """Attempts to gracefully shut down the worker thread""" if self.parent_pid != os.getpid(): - # We're not the real job queue, do nothing + # We're not the real queue, do nothing return else: log.info( "sending stop signal to worker thread" ) self.running = False -# if not self.track_jobs_in_database: -# self.queue.put( self.STOP_SIGNAL ) self.sleeper.wake() log.info( "cloud manager stopped" ) self.dispatcher.shutdown() @@ -649,11 +447,6 @@ def get_all_stores( self ): """ Returns all storage volumes' database objects associated with this UCI. """ return model.CloudStore.filter( model.CloudStore.c.uci_id == self.uci_id ).all() -# svs = model.CloudStore.filter( model.CloudStore.c.uci_id == self.uci_id ).all() -# svl = [] # storage volume list -# for sv in svs: -# svl.append( sv.volume_id ) -# return svl def get_uci( self ): """ Returns database object for given UCI. """ @@ -679,336 +472,6 @@ uci.state = 'deleted' # for bookkeeping reasons, mark as deleted but don't actually delete. uci.flush() -#class JobWrapper( object ): -# """ -# Wraps a 'model.Job' with convience methods for running processes and -# state management. -# """ -# def __init__(self, job, tool, queue ): -# self.job_id = job.id -# # This is immutable, we cache it for the scheduling policy to use if needed -# self.session_id = job.session_id -# self.tool = tool -# self.queue = queue -# self.app = queue.app -# self.extra_filenames = [] -# self.command_line = None -# self.galaxy_lib_dir = None -# # With job outputs in the working directory, we need the working -# # directory to be set before prepare is run, or else premature deletion -# # and job recovery fail. -# self.working_directory = \ -# os.path.join( self.app.config.job_working_directory, str( self.job_id ) ) -# self.output_paths = None -# self.external_output_metadata = metadata.JobExternalOutputMetadataWrapper( job ) #wrapper holding the info required to restore and clean up from files used for setting metadata externally -# -# def get_param_dict( self ): -# """ -# Restore the dictionary of parameters from the database. -# """ -# job = model.Job.get( self.job_id ) -# param_dict = dict( [ ( p.name, p.value ) for p in job.parameters ] ) -# param_dict = self.tool.params_from_strings( param_dict, self.app ) -# return param_dict -# -# def prepare( self ): -# """ -# Prepare the job to run by creating the working directory and the -# config files. -# """ -# mapping.context.current.clear() #this prevents the metadata reverting that has been seen in conjunction with the PBS job runner -# if not os.path.exists( self.working_directory ): -# os.mkdir( self.working_directory ) -# # Restore parameters from the database -# job = model.Job.get( self.job_id ) -# incoming = dict( [ ( p.name, p.value ) for p in job.parameters ] ) -# incoming = self.tool.params_from_strings( incoming, self.app ) -# # Do any validation that could not be done at job creation -# self.tool.handle_unvalidated_param_values( incoming, self.app ) -# # Restore input / output data lists -# inp_data = dict( [ ( da.name, da.dataset ) for da in job.input_datasets ] ) -# out_data = dict( [ ( da.name, da.dataset ) for da in job.output_datasets ] ) -# # These can be passed on the command line if wanted as $userId $userEmail -# if job.history.user: # check for anonymous user! -# userId = '%d' % job.history.user.id -# userEmail = str(job.history.user.email) -# else: -# userId = 'Anonymous' -# userEmail = 'Anonymous' -# incoming['userId'] = userId -# incoming['userEmail'] = userEmail -# # Build params, done before hook so hook can use -# param_dict = self.tool.build_param_dict( incoming, inp_data, out_data, self.get_output_fnames(), self.working_directory ) -# # Certain tools require tasks to be completed prior to job execution -# # ( this used to be performed in the "exec_before_job" hook, but hooks are deprecated ). -# if self.tool.tool_type is not None: -# out_data = self.tool.exec_before_job( self.queue.app, inp_data, out_data, param_dict ) -# # Run the before queue ("exec_before_job") hook -# self.tool.call_hook( 'exec_before_job', self.queue.app, inp_data=inp_data, -# out_data=out_data, tool=self.tool, param_dict=incoming) -# mapping.context.current.flush() -# # Build any required config files -# config_filenames = self.tool.build_config_files( param_dict, self.working_directory ) -# # FIXME: Build the param file (might return None, DEPRECATED) -# param_filename = self.tool.build_param_file( param_dict, self.working_directory ) -# # Build the job's command line -# self.command_line = self.tool.build_command_line( param_dict ) -# # FIXME: for now, tools get Galaxy's lib dir in their path -# if self.command_line and self.command_line.startswith( 'python' ): -# self.galaxy_lib_dir = os.path.abspath( "lib" ) # cwd = galaxy root -# # We need command_line persisted to the db in order for Galaxy to re-queue the job -# # if the server was stopped and restarted before the job finished -# job.command_line = self.command_line -# job.flush() -# # Return list of all extra files -# extra_filenames = config_filenames -# if param_filename is not None: -# extra_filenames.append( param_filename ) -# self.param_dict = param_dict -# self.extra_filenames = extra_filenames -# return extra_filenames -# -# def fail( self, message, exception=False ): -# """ -# Indicate job failure by setting state and message on all output -# datasets. -# """ -# job = model.Job.get( self.job_id ) -# job.refresh() -# # if the job was deleted, don't fail it -# if not job.state == model.Job.states.DELETED: -# # Check if the failure is due to an exception -# if exception: -# # Save the traceback immediately in case we generate another -# # below -# job.traceback = traceback.format_exc() -# # Get the exception and let the tool attempt to generate -# # a better message -# etype, evalue, tb = sys.exc_info() -# m = self.tool.handle_job_failure_exception( evalue ) -# if m: -# message = m -# if self.app.config.outputs_to_working_directory: -# for dataset_path in self.get_output_fnames(): -# try: -# shutil.move( dataset_path.false_path, dataset_path.real_path ) -# log.debug( "fail(): Moved %s to %s" % ( dataset_path.false_path, dataset_path.real_path ) ) -# except ( IOError, OSError ), e: -# log.error( "fail(): Missing output file in working directory: %s" % e ) -# for dataset_assoc in job.output_datasets: -# dataset = dataset_assoc.dataset -# dataset.refresh() -# dataset.state = dataset.states.ERROR -# dataset.blurb = 'tool error' -# dataset.info = message -# dataset.set_size() -# dataset.flush() -# job.state = model.Job.states.ERROR -# job.command_line = self.command_line -# job.info = message -# job.flush() -# # If the job was deleted, just clean up -# self.cleanup() -# -# def change_state( self, state, info = False ): -# job = model.Job.get( self.job_id ) -# job.refresh() -# for dataset_assoc in job.output_datasets: -# dataset = dataset_assoc.dataset -# dataset.refresh() -# dataset.state = state -# if info: -# dataset.info = info -# dataset.flush() -# if info: -# job.info = info -# job.state = state -# job.flush() -# -# def get_state( self ): -# job = model.Job.get( self.job_id ) -# job.refresh() -# return job.state -# -# def set_runner( self, runner_url, external_id ): -# job = model.Job.get( self.job_id ) -# job.refresh() -# job.job_runner_name = runner_url -# job.job_runner_external_id = external_id -# job.flush() -# -# def finish( self, stdout, stderr ): -# """ -# Called to indicate that the associated command has been run. Updates -# the output datasets based on stderr and stdout from the command, and -# the contents of the output files. -# """ -# # default post job setup -# mapping.context.current.clear() -# job = model.Job.get( self.job_id ) -# # if the job was deleted, don't finish it -# if job.state == job.states.DELETED: -# self.cleanup() -# return -# elif job.state == job.states.ERROR: -# # Job was deleted by an administrator -# self.fail( job.info ) -# return -# if stderr: -# job.state = "error" -# else: -# job.state = 'ok' -# if self.app.config.outputs_to_working_directory: -# for dataset_path in self.get_output_fnames(): -# try: -# shutil.move( dataset_path.false_path, dataset_path.real_path ) -# log.debug( "finish(): Moved %s to %s" % ( dataset_path.false_path, dataset_path.real_path ) ) -# except ( IOError, OSError ): -# self.fail( "Job %s's output dataset(s) could not be read" % job.id ) -# return -# for dataset_assoc in job.output_datasets: -# #should this also be checking library associations? - can a library item be added from a history before the job has ended? - lets not allow this to occur -# for dataset in dataset_assoc.dataset.dataset.history_associations: #need to update all associated output hdas, i.e. history was shared with job running -# dataset.blurb = 'done' -# dataset.peek = 'no peek' -# dataset.info = stdout + stderr -# dataset.set_size() -# if stderr: -# dataset.blurb = "error" -# elif dataset.has_data(): -# #if a dataset was copied, it won't appear in our dictionary: -# #either use the metadata from originating output dataset, or call set_meta on the copies -# #it would be quicker to just copy the metadata from the originating output dataset, -# #but somewhat trickier (need to recurse up the copied_from tree), for now we'll call set_meta() -# if not self.external_output_metadata.external_metadata_set_successfully( dataset ): -# # Only set metadata values if they are missing... -# dataset.set_meta( overwrite = False ) -# else: -# #load metadata from file -# #we need to no longer allow metadata to be edited while the job is still running, -# #since if it is edited, the metadata changed on the running output will no longer match -# #the metadata that was stored to disk for use via the external process, -# #and the changes made by the user will be lost, without warning or notice -# dataset.metadata.from_JSON_dict( self.external_output_metadata.get_output_filenames_by_dataset( dataset ).filename_out ) -# if self.tool.is_multi_byte: -# dataset.set_multi_byte_peek() -# else: -# dataset.set_peek() -# else: -# dataset.blurb = "empty" -# dataset.flush() -# if stderr: -# dataset_assoc.dataset.dataset.state = model.Dataset.states.ERROR -# else: -# dataset_assoc.dataset.dataset.state = model.Dataset.states.OK -# dataset_assoc.dataset.dataset.flush() -# -# # Save stdout and stderr -# if len( stdout ) > 32768: -# log.error( "stdout for job %d is greater than 32K, only first part will be logged to database" % job.id ) -# job.stdout = stdout[:32768] -# if len( stderr ) > 32768: -# log.error( "stderr for job %d is greater than 32K, only first part will be logged to database" % job.id ) -# job.stderr = stderr[:32768] -# # custom post process setup -# inp_data = dict( [ ( da.name, da.dataset ) for da in job.input_datasets ] ) -# out_data = dict( [ ( da.name, da.dataset ) for da in job.output_datasets ] ) -# param_dict = dict( [ ( p.name, p.value ) for p in job.parameters ] ) # why not re-use self.param_dict here? ##dunno...probably should, this causes tools.parameters.basic.UnvalidatedValue to be used in following methods instead of validated and transformed values during i.e. running workflows -# param_dict = self.tool.params_from_strings( param_dict, self.app ) -# # Check for and move associated_files -# self.tool.collect_associated_files(out_data, self.working_directory) -# # Create generated output children and primary datasets and add to param_dict -# collected_datasets = {'children':self.tool.collect_child_datasets(out_data),'primary':self.tool.collect_primary_datasets(out_data)} -# param_dict.update({'__collected_datasets__':collected_datasets}) -# # Certain tools require tasks to be completed after job execution -# # ( this used to be performed in the "exec_after_process" hook, but hooks are deprecated ). -# if self.tool.tool_type is not None: -# self.tool.exec_after_process( self.queue.app, inp_data, out_data, param_dict, job = job ) -# # Call 'exec_after_process' hook -# self.tool.call_hook( 'exec_after_process', self.queue.app, inp_data=inp_data, -# out_data=out_data, param_dict=param_dict, -# tool=self.tool, stdout=stdout, stderr=stderr ) -# # TODO -# # validate output datasets -# job.command_line = self.command_line -# mapping.context.current.flush() -# log.debug( 'job %d ended' % self.job_id ) -# self.cleanup() -# -# def cleanup( self ): -# # remove temporary files -# try: -# for fname in self.extra_filenames: -# os.remove( fname ) -# if self.working_directory is not None: -# shutil.rmtree( self.working_directory ) -# if self.app.config.set_metadata_externally: -# self.external_output_metadata.cleanup_external_metadata() -# except: -# log.exception( "Unable to cleanup job %d" % self.job_id ) -# -# def get_command_line( self ): -# return self.command_line -# -# def get_session_id( self ): -# return self.session_id -# -# def get_input_fnames( self ): -# job = model.Job.get( self.job_id ) -# filenames = [] -# for da in job.input_datasets: #da is JobToInputDatasetAssociation object -# if da.dataset: -# filenames.append( da.dataset.file_name ) -# #we will need to stage in metadata file names also -# #TODO: would be better to only stage in metadata files that are actually needed (found in command line, referenced in config files, etc.) -# for key, value in da.dataset.metadata.items(): -# if isinstance( value, model.MetadataFile ): -# filenames.append( value.file_name ) -# return filenames -# -# def get_output_fnames( self ): -# if self.output_paths is not None: -# return self.output_paths -# -# class DatasetPath( object ): -# def __init__( self, real_path, false_path = None ): -# self.real_path = real_path -# self.false_path = false_path -# def __str__( self ): -# if self.false_path is None: -# return self.real_path -# else: -# return self.false_path -# -# job = model.Job.get( self.job_id ) -# if self.app.config.outputs_to_working_directory: -# self.output_paths = [] -# for name, data in [ ( da.name, da.dataset.dataset ) for da in job.output_datasets ]: -# false_path = os.path.abspath( os.path.join( self.working_directory, "galaxy_dataset_%d.dat" % data.id ) ) -# self.output_paths.append( DatasetPath( data.file_name, false_path ) ) -# else: -# self.output_paths = [ DatasetPath( da.dataset.file_name ) for da in job.output_datasets ] -# return self.output_paths -# -# def check_output_sizes( self ): -# sizes = [] -# output_paths = self.get_output_fnames() -# for outfile in [ str( o ) for o in output_paths ]: -# sizes.append( ( outfile, os.stat( outfile ).st_size ) ) -# return sizes -# def setup_external_metadata( self, exec_dir = None, tmp_dir = None, dataset_files_path = None, config_root = None, datatypes_config = None, **kwds ): -# if tmp_dir is None: -# #this dir should should relative to the exec_dir -# tmp_dir = self.app.config.new_file_path -# if dataset_files_path is None: -# dataset_files_path = self.app.model.Dataset.file_path -# if config_root is None: -# config_root = self.app.config.root -# if datatypes_config is None: -# datatypes_config = self.app.config.datatypes_config -# job = model.Job.get( self.job_id ) -# return self.external_output_metadata.setup_external_metadata( [ output_dataset_assoc.dataset for output_dataset_assoc in job.output_datasets ], exec_dir = exec_dir, tmp_dir = tmp_dir, dataset_files_path = dataset_files_path, config_root = config_root, datatypes_config = datatypes_config, **kwds ) - class CloudProvider( object ): def __init__( self, app ): import providers.eucalyptus @@ -1018,63 +481,11 @@ self.cloud_provider = {} self.cloud_provider["eucalyptus"] = providers.eucalyptus.EucalyptusCloudProvider( app ) self.cloud_provider["ec2"] = providers.ec2.EC2CloudProvider( app ) - -# start_cloud_provider = None -# if app.config.start_job_runners is not None: -# start_cloud_provider.extend( app.config.start_job_runners.split(",") ) -# for provider_name in start_cloud_provider: -# self.provider_name = app.config.cloud_provider -# if self.provider_name == "eucalyptus": -# import providers.eucalyptus -# self.cloud_provider[self.provider_name] = providers.eucalyptus.EucalyptusCloudProvider( app ) -# elif self.provider_name == "ec2": -# import providers.ec2 -# self.cloud_provider[self.provider_name] = providers.ec2.EC2CloudProvider( app ) -# else: -# log.error( "Unable to start unknown cloud provider: %s" %self.provider_name ) - + def put( self, uci_wrapper ): """ Put given request for UCI manipulation into provider's request queue.""" -# log.debug( "Adding UCI '%s' manipulation request into cloud manager's queue." % uci_wrapper.name ) self.cloud_provider[uci_wrapper.get_provider_type()].put( uci_wrapper ) - - - - def createUCI( self, uci ): - """ - Createse User Configured Instance (UCI). Essentially, creates storage volume. - """ - log.debug( "Creating UCI '%s'" % uci.name ) - self.cloud_provider[self.provider_name].createUCI( uci ) - - def deleteUCI( self, uci ): - """ - Deletes UCI. NOTE that this implies deletion of any and all data associated - with this UCI from the cloud. All data will be deleted. - """ - log.debug( "Deleting UCI '%s'" % uci.name ) - self.cloud_provider[self.provider_name].deleteUCI( uci ) - - def addStorageToUCI( self, uci ): - """ Adds more storage to specified UCI """ - - def startUCI( self, uci ): - """ - Starts an instance of named UCI on the cloud. This implies, mounting of - storage and starting Galaxy instance. - """ - log.debug( "Starting UCI '%s'" % uci.name ) - self.cloud_provider[self.provider_name].startUCI( uci ) - - def stopUCI( self, uci ): - """ - Stops cloud instance associated with named UCI. This also implies - stopping of Galaxy and unmounting of the file system. - """ - log.debug( "Stopping UCI '%s'" % uci.name ) - self.cloud_provider[self.provider_name].stopUCI( uci ) - def update( self ): """ Runs a global status update across all providers for all UCIs in state other than 'terminated' and 'available'. @@ -1084,102 +495,10 @@ # log.debug( "Running global update for provider: '%s'" % provider ) self.cloud_provider[provider].update() - def recover( self, job, job_wrapper ): - runner_name = ( job.job_runner_name.split(":", 1) )[0] - log.debug( "recovering job %d in %s runner" %( job.id, runner_name ) ) - self.cloud_provider[runner_name].recover( job, job_wrapper ) - def shutdown( self ): for runner in self.cloud_provider.itervalues(): runner.shutdown() -class JobStopQueue( object ): - """ - A queue for jobs which need to be terminated prematurely. - """ - STOP_SIGNAL = object() - def __init__( self, app, dispatcher ): - self.app = app - self.dispatcher = dispatcher - - # Keep track of the pid that started the job manager, only it - # has valid threads - self.parent_pid = os.getpid() - # Contains new jobs. Note this is not used if track_jobs_in_database is True - self.queue = Queue() - - # Contains jobs that are waiting (only use from monitor thread) - self.waiting = [] - - # Helper for interruptable sleep - self.sleeper = Sleeper() - self.running = True - self.monitor_thread = threading.Thread( target=self.monitor ) - self.monitor_thread.start() - log.info( "job stopper started" ) - - def monitor( self ): - """ - Continually iterate the waiting jobs, stop any that are found. - """ - # HACK: Delay until after forking, we need a way to do post fork notification!!! - time.sleep( 10 ) - while self.running: - try: - self.monitor_step() - except: - log.exception( "Exception in monitor_step" ) - # Sleep - self.sleeper.sleep( 1 ) - - def monitor_step( self ): - """ - Called repeatedly by `monitor` to stop jobs. - """ - # Pull all new jobs from the queue at once - jobs = [] - try: - while 1: - ( job_id, error_msg ) = self.queue.get_nowait() - if job_id is self.STOP_SIGNAL: - return - # Append to watch queue - jobs.append( ( job_id, error_msg ) ) - except Empty: - pass - - for job_id, error_msg in jobs: - job = model.Job.get( job_id ) - job.refresh() - # if desired, error the job so we can inform the user. - if error_msg is not None: - job.state = job.states.ERROR - job.info = error_msg - else: - job.state = job.states.DELETED - job.flush() - # if job is in JobQueue or FooJobRunner's put method, - # job_runner_name will be unset and the job will be dequeued due to - # state change above - if job.job_runner_name is not None: - # tell the dispatcher to stop the job - self.dispatcher.stop( job ) - - def put( self, job_id, error_msg=None ): - self.queue.put( ( job_id, error_msg ) ) - - def shutdown( self ): - """Attempts to gracefully shut down the worker thread""" - if self.parent_pid != os.getpid(): - # We're not the real job queue, do nothing - return - else: - log.info( "sending stop signal to worker thread" ) - self.running = False - self.queue.put( ( self.STOP_SIGNAL, None ) ) - self.sleeper.wake() - log.info( "job stopper stopped" ) - class NoopCloudMonitor( object ): """ Implements the CloudMonitor interface but does nothing diff -r 022ac0f64679 -r be631ed97541 lib/galaxy/cloud/providers/ec2.py --- a/lib/galaxy/cloud/providers/ec2.py Wed Oct 28 16:35:36 2009 -0400 +++ b/lib/galaxy/cloud/providers/ec2.py Thu Oct 29 11:46:43 2009 -0400 @@ -56,7 +56,6 @@ self.key_pair = "galaxy-keypair" self.queue = Queue() - #TODO: Use multiple threads to process requests? self.threads = [] nworkers = 5 log.info( "Starting EC2 cloud controller workers" ) @@ -72,19 +71,17 @@ while 1: uci_wrapper = self.queue.get() -# uci = uci_wrapper.get_uci() - log.debug( '[%d] uci type: %s' % ( cnt, uci_wrapper.get_name() ) ) uci_state = uci_wrapper.get_state() if uci_state is self.STOP_SIGNAL: return try: - if uci_state==uci_states.NEW: # "new": + if uci_state==uci_states.NEW: self.createUCI( uci_wrapper ) - elif uci_state==uci_states.DELETING: #"deleting": + elif uci_state==uci_states.DELETING: self.deleteUCI( uci_wrapper ) - elif uci_state==uci_states.SUBMITTED: #"submitted": + elif uci_state==uci_states.SUBMITTED: self.startUCI( uci_wrapper ) - elif uci_state==uci_states.SHUTTING_DOWN: #"shutting-down": + elif uci_state==uci_states.SHUTTING_DOWN: self.stopUCI( uci_wrapper ) except: log.exception( "Uncaught exception executing request." ) @@ -113,7 +110,6 @@ try: kp = conn.get_key_pair( self.key_pair ) for inst in instances: -# log.debug("inst: '%s'" % inst ) uci_wrapper.set_key_pair( inst, kp.name ) return kp.name except boto.exception.EC2ResponseError, e: # No keypair under this name exists so create it @@ -135,22 +131,7 @@ For valid sizes, see http://aws.amazon.com/ec2/instance-types/ """ return model.CloudImage.filter( model.CloudImage.table.c.id==2 ).first().image_id - -# def get_instances( self, uci ): -# """ -# Get objects of instances that are pending or running and are connected to uci object -# """ -# instances = trans.sa_session.query( model.CloudInstance ) \ -# .filter_by( user=user, uci_id=uci.id ) \ -# .filter( or_(model.CloudInstance.table.c.state=="running", model.CloudInstance.table.c.state=="pending" ) ) \ -# .first() -# #.all() #TODO: return all but need to edit calling method(s) to handle list -# -# instances = uci.instance -# -# return instances - - + def shutdown( self ): """Attempts to gracefully shut down the monitor thread""" log.info( "sending stop signal to worker threads in EC2 cloud manager" ) @@ -170,7 +151,6 @@ and registers relevant information in Galaxy database. """ conn = self.get_connection( uci_wrapper ) - # Temporary code - need to ensure user selects zone at UCI creation time! if uci_wrapper.get_uci_availability_zone()=='': log.info( "Availability zone for UCI (i.e., storage volume) was not selected, using default zone: %s" % self.zone ) uci_wrapper.set_store_availability_zone( self.zone ) @@ -262,12 +242,11 @@ for i_index in i_indexes: mi_id = self.get_mi_id( uci_wrapper.get_type( i_index ) ) -# log.debug( "mi_id: %s, uci_wrapper.get_key_pair_name( i_index ): %s" % ( mi_id, uci_wrapper.get_key_pair_name( i_index ) ) ) uci_wrapper.set_mi( i_index, mi_id ) # Check if galaxy security group exists (and create it if it does not) -# log.debug( '***** Setting up security group' ) security_group = 'galaxyWeb' + log.debug( "Setting up '%s' security group." % security_group ) sgs = conn.get_all_security_groups() # security groups gsgt = False # galaxy security group test for sg in sgs: @@ -303,39 +282,6 @@ uci_wrapper.change_state( s, i_id, s ) log.debug( "Instance of UCI '%s' started, current state: '%s'" % ( uci_wrapper.get_name(), uci_wrapper.get_state() ) ) - - -# # Wait until instance gets running and then update the DB -# while s!="running": -# log.debug( "Waiting on instance '%s' to start up (reservation ID: %s); current state: %s" % ( uci.instance[0].instance_id, uci.instance[0].reservation_id, s ) ) -# time.sleep( 15 ) -# s = reservation.i_indexes[0].update() -# -# # Update instance data in local DB -# uci.instance[0].state = s -# uci.instance[0].public_dns = reservation.i_indexes[0].dns_name -# uci.instance[0].private_dns = reservation.i_indexes[0].private_dns_name -# uci.instance[0].flush() -# # Update storage data in local DB w/ volume state info. NOTE that this only captures current volume state -# # and does not connect or wait on connection between instance and volume to be established -# vl = model.CloudStore.filter( model.CloudStore.c.uci_id == uci.id ).all() -# vols = [] -# for v in vl: -# vols.append( v.volume_id ) -# try: -# volumes = conn.get_all_volumes( vols ) -# for i, v in enumerate( volumes ): -# uci.store[i].i_id = v.instance_id -# uci.store[i].status = v.status -# uci.store[i].device = v.device -# uci.store[i].flush() -# except BotoServerError: -# log.debug( "Error getting volume(s) attached to instance. Volume status was not updated." ) -# -# uci.state = s -# uci.flush() - - def stopUCI( self, uci_wrapper): """ Stops all of cloud instances associated with given UCI. @@ -344,14 +290,8 @@ # Get all instances associated with given UCI il = uci_wrapper.get_instances_ids() # instance list -# log.debug( 'List of instances being terminated: %s' % il ) rl = conn.get_all_instances( il ) # Reservation list associated with given instances - -# tState = conn.terminate_instances( il ) -# # TODO: Need to update instance stop time (for all individual instances) -# stop_time = datetime.utcnow() -# uci_wrapper.set_stop_time( stop_time ) - + # Initiate shutdown of all instances under given UCI cnt = 0 stopped = [] @@ -364,26 +304,7 @@ uci_wrapper.change_state( instance_id=inst.id, i_state=inst.update() ) stopped.append( inst ) -# uci_wrapper.change_state( uci_state='available' ) uci_wrapper.reset_uci_launch_time() - -# # Wait for all instances to actually terminate and update local DB -# terminated=0 -# while terminated!=len( rl ): -# for i, r in enumerate( rl ): -# log.debug( "r state: %s" % r.instances[0].state ) -# state = r.instances[0].update() -# if state=='terminated': -# uci.instance[i].state = state -# uci.instance[i].flush() -# terminated += 1 -# time.sleep ( 5 ) -# -# # Reset UCI state -# uci.state = 'available' -# uci.launch_time = None -# uci.flush() -# log.debug( "Termination was initiated for all instances of UCI '%s'." % uci_wrapper.get_name() ) diff -r 022ac0f64679 -r be631ed97541 lib/galaxy/cloud/providers/eucalyptus.py --- a/lib/galaxy/cloud/providers/eucalyptus.py Wed Oct 28 16:35:36 2009 -0400 +++ b/lib/galaxy/cloud/providers/eucalyptus.py Thu Oct 29 11:46:43 2009 -0400 @@ -56,7 +56,6 @@ self.key_pair = "galaxy-keypair" self.queue = Queue() - #TODO: Use multiple threads to process requests? self.threads = [] nworkers = 5 log.info( "Starting eucalyptus cloud controller workers" ) @@ -72,19 +71,18 @@ while 1: uci_wrapper = self.queue.get() -# uci = uci_wrapper.get_uci() log.debug( '[%d] uci type: %s' % ( cnt, uci_wrapper.get_name() ) ) uci_state = uci_wrapper.get_state() if uci_state is self.STOP_SIGNAL: return try: - if uci_state==uci_states.NEW: # "new": + if uci_state==uci_states.NEW: self.createUCI( uci_wrapper ) - elif uci_state==uci_states.DELETING: #"deleting": + elif uci_state==uci_states.DELETING: self.deleteUCI( uci_wrapper ) - elif uci_state==uci_states.SUBMITTED: #"submitted": + elif uci_state==uci_states.SUBMITTED: self.startUCI( uci_wrapper ) - elif uci_state==uci_states.SHUTTING_DOWN: #"shutting-down": + elif uci_state==uci_states.SHUTTING_DOWN: self.stopUCI( uci_wrapper ) except: log.exception( "Uncaught exception executing request." ) @@ -134,22 +132,7 @@ """ log.debug( "image id: '%s'" % model.CloudImage.get( 1 ).image_id ) return model.CloudImage.get( 1 ).image_id - -# def get_instances( self, uci ): -# """ -# Get objects of instances that are pending or running and are connected to uci object -# """ -# instances = trans.sa_session.query( model.CloudInstance ) \ -# .filter_by( user=user, uci_id=uci.id ) \ -# .filter( or_(model.CloudInstance.table.c.state=="running", model.CloudInstance.table.c.state=="pending" ) ) \ -# .first() -# #.all() #TODO: return all but need to edit calling method(s) to handle list -# -# instances = uci.instance -# -# return instances - - + def shutdown( self ): """Attempts to gracefully shut down the monitor thread""" log.info( "sending stop signal to worker threads in eucalyptus cloud manager" ) @@ -169,7 +152,6 @@ and registers relevant information in Galaxy database. """ conn = self.get_connection( uci_wrapper ) - # Temporary code - need to ensure user selects zone at UCI creation time! if uci_wrapper.get_uci_availability_zone()=='': log.info( "Availability zone for UCI (i.e., storage volume) was not selected, using default zone: %s" % self.zone ) uci_wrapper.set_store_availability_zone( self.zone ) @@ -245,21 +227,7 @@ mi_id = self.get_mi_id( uci_wrapper.get_type( i_index ) ) log.debug( "mi_id: %s, uci_wrapper.get_key_pair_name( i_index ): %s" % ( mi_id, uci_wrapper.get_key_pair_name( i_index ) ) ) uci_wrapper.set_mi( i_index, mi_id ) - - # log.debug( '***** Setting up security group' ) - # If not existent, setup galaxy security group - # try: - # gSecurityGroup = conn.create_security_group('galaxy', 'Security group for Galaxy.') - # gSecurityGroup.authorize( 'tcp', 80, 80, '0.0.0.0/0' ) # Open HTTP port - # gSecurityGroup.authorize( 'tcp', 22, 22, '0.0.0.0/0' ) # Open SSH port - # except: - # pass - # sgs = conn.get_all_security_groups() - # for i in range( len( sgs ) ): - # if sgs[i].name == "galaxy": - # sg.append( sgs[i] ) - # break # only 1 security group w/ this type can exist, so continue - + log.debug( "***** Starting UCI instance '%s'" % uci_wrapper.get_name() ) log.debug( 'Using following command: conn.run_instances( image_id=%s, key_name=%s )' % ( mi_id, uci_wrapper.get_key_pair_name( i_index ) ) ) reservation = conn.run_instances( image_id=mi_id, key_name=uci_wrapper.get_key_pair_name( i_index ) ) @@ -276,39 +244,6 @@ uci_wrapper.change_state( s, i_id, s ) log.debug( "Instance of UCI '%s' started, current state: '%s'" % ( uci_wrapper.get_name(), uci_wrapper.get_state() ) ) - - -# # Wait until instance gets running and then update the DB -# while s!="running": -# log.debug( "Waiting on instance '%s' to start up (reservation ID: %s); current state: %s" % ( uci.instance[0].instance_id, uci.instance[0].reservation_id, s ) ) -# time.sleep( 15 ) -# s = reservation.i_indexes[0].update() -# -# # Update instance data in local DB -# uci.instance[0].state = s -# uci.instance[0].public_dns = reservation.i_indexes[0].dns_name -# uci.instance[0].private_dns = reservation.i_indexes[0].private_dns_name -# uci.instance[0].flush() -# # Update storage data in local DB w/ volume state info. NOTE that this only captures current volume state -# # and does not connect or wait on connection between instance and volume to be established -# vl = model.CloudStore.filter( model.CloudStore.c.uci_id == uci.id ).all() -# vols = [] -# for v in vl: -# vols.append( v.volume_id ) -# try: -# volumes = conn.get_all_volumes( vols ) -# for i, v in enumerate( volumes ): -# uci.store[i].i_id = v.instance_id -# uci.store[i].status = v.status -# uci.store[i].device = v.device -# uci.store[i].flush() -# except BotoServerError: -# log.debug( "Error getting volume(s) attached to instance. Volume status was not updated." ) -# -# uci.state = s -# uci.flush() - - def stopUCI( self, uci_wrapper): """ Stops all of cloud instances associated with given UCI. @@ -332,30 +267,9 @@ uci_wrapper.change_state( instance_id=inst.id, i_state=inst.update() ) stopped.append( inst ) -# uci_wrapper.change_state( uci_state='available' ) uci_wrapper.reset_uci_launch_time() - -# # Wait for all instances to actually terminate and update local DB -# terminated=0 -# while terminated!=len( rl ): -# for i, r in enumerate( rl ): -# log.debug( "r state: %s" % r.instances[0].state ) -# state = r.instances[0].update() -# if state=='terminated': -# uci.instance[i].state = state -# uci.instance[i].flush() -# terminated += 1 -# time.sleep ( 5 ) -# -# # Reset UCI state -# uci.state = 'available' -# uci.launch_time = None -# uci.flush() -# log.debug( "Termination was initiated for all instances of UCI '%s'." % uci_wrapper.get_name() ) - - # dbInstances = get_instances( trans, uci ) #TODO: handle list! # # # Get actual cloud instance object @@ -377,7 +291,6 @@ # store.i_id = None # store.status = volStat # log.debug ( '***** volume status: %s' % volStat ) -# # # # Stop the instance and update status in local database # cloudInstance.stop() @@ -495,7 +408,6 @@ region=euca_region, path=uci.credentials.provider.path )# Get reservations handle for given store vl = conn.get_all_volumes( [store.volume_id] ) -# log.debug( "Store '%s' vl: '%s'" % ( store.volume_id, vl ) ) # Update store status in local DB with info from cloud provider if store.status != vl[0].status: # In case something failed during creation of UCI but actual storage volume was created and yet diff -r 022ac0f64679 -r be631ed97541 lib/galaxy/web/controllers/cloud.py --- a/lib/galaxy/web/controllers/cloud.py Wed Oct 28 16:35:36 2009 -0400 +++ b/lib/galaxy/web/controllers/cloud.py Thu Oct 29 11:46:43 2009 -0400 @@ -12,7 +12,6 @@ from galaxy.util.odict import odict from galaxy.util.bunch import Bunch from galaxy.util.topsort import topsort, topsort_levels, CycleError -from galaxy.workflow.modules import * from galaxy.model.mapping import desc from galaxy.model.orm import * from datetime import datetime, timedelta @@ -69,12 +68,6 @@ Render cloud main page (management of cloud resources) """ user = trans.get_user() -# pendingInstances = trans.sa_session.query( model.UCI ) \ -# .filter_by( user=user, state="pending" ) \ -# .all() -# -# for i inupdate_in range( len ( pendingInstances ) ): -# stance_state( trans, pendingInstances[i].id ) cloudCredentials = trans.sa_session.query( model.CloudUserCredentials ) \ .filter_by( user=user ) \ @@ -104,7 +97,6 @@ .all() # Check after update there are instances in pending state; if so, display message - # TODO: Auto-refresh once instance is running pendingInstances = trans.sa_session.query( model.UCI ) \ .filter_by( user=user ) \ .filter( or_( model.UCI.c.state==uci_states.PENDING, #"pending" , \ @@ -128,7 +120,6 @@ liveInstances = liveInstances, prevInstances = prevInstances ) - @web.expose @web.require_login( "use Galaxy cloud" ) def makeDefault( self, trans, id=None ): """ @@ -337,8 +328,6 @@ inst_error = "No registered cloud images. You must contact administrator to add some before proceeding." log.debug("AttributeError: %s " % str( ae ) ) - #TODO: based on user credentials (i.e., provider) selected, zone options will be different (e.g., EC2: us-east-1a vs EPC: epc) - return trans.fill_template( "cloud/configure_uci.mako", instanceName = instanceName, credName = storedCreds, @@ -427,8 +416,6 @@ session.flush() # Log and display the management page trans.set_message( "Credential '%s' edited." % credentials.name ) -# if defaultCred: -# self.makeDefault( trans, credentials.id) return self.list( trans ) @web.expose @@ -507,8 +494,6 @@ # Log and display the management page trans.log_event( "User added new credentials" ) trans.set_message( "Credential '%s' created" % credentials.name ) -# if defaultCred: -# self.makeDefault( trans, credentials.id) return self.list( trans ) providers = trans.sa_session.query( model.CloudProvider ).filter_by( user=user ).all() @@ -521,13 +506,6 @@ providers = providers ) -# return trans.show_form( -# web.FormBuilder( web.url_for(), "Add credentials", submit_text="Add" ) -# .add_text( "credName", "Credentials name", value="Unnamed credentials", error=cred_error ) -# .add_text( "providerName", "Cloud provider name", value="ec2 or eucalyptus", error=provider_error ) -# .add_text( "accessKey", "Access key", value='', error=accessKey_error ) -# .add_password( "secretKey", "Secret key", value='', error=secretKey_error ) ) - @web.expose @web.require_login( "view credentials" ) def view( self, trans, id=None ): @@ -770,18 +748,6 @@ # Looks good return stored -def get_default_credentials( trans, check_ownership=True ): - """ - Get a StoredUserCredntials from the database by 'default' setting, verifying ownership. - """ - user = trans.get_user() - # Load credentials from database - stored = trans.sa_session.query( model.CloudUserCredentials ) \ - .filter_by (user=user, defaultCred=True) \ - .first() - - return stored - def get_uci( trans, id, check_ownership=True ): """ Get a UCI object from the database by id, verifying ownership. @@ -891,204 +857,3 @@ # Actually, probably return key_pair to calling method and store name & key from there... return key_pair.name - -def update_instance_state( trans, id ): - """ - Update state of instances associated with given UCI id and store state in local database. Also update - state of the given UCI. - """ - uci = get_uci( trans, id ) - # Get list of instances associated with given uci as they are stored in local database - dbInstances = get_instances( trans, uci ) # TODO: handle list (currently only 1 instance can correspond to 1 UCI) - oldState = dbInstances.state - # Establish connection with cloud - conn = get_connection( trans ) - # Get actual cloud instance object - cloudInstance = get_cloud_instance( conn, dbInstances.instance_id ) - # Update instance status - cloudInstance.update() - dbInstances.state = cloudInstance.state - log.debug( "Updating instance %s state; current state: %s" % ( str( cloudInstance ).split(":")[1], cloudInstance.state ) ) - # Update state of UCI (TODO: once more than 1 instance is assoc. w/ 1 UCI, this will be need to be updated differently) - uci.state = dbInstances.state - # Persist - session = trans.sa_session - session.save_or_update( dbInstances ) - session.save_or_update( uci ) - session.flush() - - # If instance is now running, update/process instance (i.e., mount file system, start Galaxy, update DB with DNS) - if oldState==instance_states.PENDING and dbInstances.state==instance_states.RUNNING: - update_instance( trans, dbInstances, cloudInstance, conn, uci ) - - -def update_instance( trans, dbInstance, cloudInstance, conn, uci ): - """ - Update instance: connect EBS volume, mount file system, start Galaxy, and update local DB w/ DNS info - - Keyword arguments: - trans -- current transaction - dbInstance -- object of 'instance' as it is stored in local database - cloudInstance -- object of 'instance' as it resides in the cloud. Functions supported by the cloud API can be - instantiated directly on this object. - conn -- cloud connection object - uci -- UCI object - """ - dbInstance.public_dns = cloudInstance.dns_name - dbInstance.private_dns = cloudInstance.private_dns_name - - # Attach storage volume(s) to instance - stores = get_stores( trans, uci ) - for i, store in enumerate( stores ): - log.debug( "Attaching volume '%s' to instance '%s'." % ( store.volume_id, dbInstance.instance_id ) ) - mntDevice = '/dev/sdb'+str(i) - volStat = conn.attach_volume( store.volume_id, dbInstance.instance_id, mntDevice ) - store.attach_time = datetime.utcnow() - store.device = mntDevice - store.i_id = dbInstance.instance_id - store.status = volStat - log.debug ( '***** volume status: %s' % volStat ) - - # Wait until instances have attached and add file system - - - - # TODO: mount storage through ZFS - # TODO: start Galaxy - - # Persist - session = trans.sa_session - session.save_or_update( dbInstance ) - session.flush() - -def attach_ordered_steps( workflow, steps ): - ordered_steps = order_workflow_steps( steps ) - if ordered_steps: - workflow.has_cycles = False - for i, step in enumerate( ordered_steps ): - step.order_index = i - workflow.steps.append( step ) - else: - workflow.has_cycles = True - workflow.steps = steps - -def edgelist_for_workflow_steps( steps ): - """ - Create a list of tuples representing edges between `WorkflowSteps` based - on associated `WorkflowStepConnection`s - """ - edges = [] - steps_to_index = dict( ( step, i ) for i, step in enumerate( steps ) ) - for step in steps: - edges.append( ( steps_to_index[step], steps_to_index[step] ) ) - for conn in step.input_connections: - edges.append( ( steps_to_index[conn.output_step], steps_to_index[conn.input_step] ) ) - return edges - -def order_workflow_steps( steps ): - """ - Perform topological sort of the steps, return ordered or None - """ - try: - edges = edgelist_for_workflow_steps( steps ) - node_order = topsort( edges ) - return [ steps[i] for i in node_order ] - except CycleError: - return None - -def order_workflow_steps_with_levels( steps ): - try: - return topsort_levels( edgelist_for_workflow_steps( steps ) ) - except CycleError: - return None - -class FakeJob( object ): - """ - Fake job object for datasets that have no creating_job_associations, - they will be treated as "input" datasets. - """ - def __init__( self, dataset ): - self.is_fake = True - self.id = "fake_%s" % dataset.id - -def get_job_dict( trans ): - """ - Return a dictionary of Job -> [ Dataset ] mappings, for all finished - active Datasets in the current history and the jobs that created them. - """ - history = trans.get_history() - # Get the jobs that created the datasets - warnings = set() - jobs = odict() - for dataset in history.active_datasets: - # FIXME: Create "Dataset.is_finished" - if dataset.state in ( 'new', 'running', 'queued' ): - warnings.add( "Some datasets still queued or running were ignored" ) - continue - - #if this hda was copied from another, we need to find the job that created the origial hda - job_hda = dataset - while job_hda.copied_from_history_dataset_association: - job_hda = job_hda.copied_from_history_dataset_association - - if not job_hda.creating_job_associations: - jobs[ FakeJob( dataset ) ] = [ ( None, dataset ) ] - - for assoc in job_hda.creating_job_associations: - job = assoc.job - if job in jobs: - jobs[ job ].append( ( assoc.name, dataset ) ) - else: - jobs[ job ] = [ ( assoc.name, dataset ) ] - return jobs, warnings - -def cleanup_param_values( inputs, values ): - """ - Remove 'Data' values from `param_values`, along with metadata cruft, - but track the associations. - """ - associations = [] - names_to_clean = [] - # dbkey is pushed in by the framework - if 'dbkey' in values: - del values['dbkey'] - root_values = values - # Recursively clean data inputs and dynamic selects - def cleanup( prefix, inputs, values ): - for key, input in inputs.items(): - if isinstance( input, ( SelectToolParameter, DrillDownSelectToolParameter ) ): - if input.is_dynamic: - values[key] = UnvalidatedValue( values[key] ) - if isinstance( input, DataToolParameter ): - tmp = values[key] - values[key] = None - # HACK: Nested associations are not yet working, but we - # still need to clean them up so we can serialize - # if not( prefix ): - if tmp: #this is false for a non-set optional dataset - associations.append( ( tmp.hid, prefix + key ) ) - # Cleanup the other deprecated crap associated with datasets - # as well. Worse, for nested datasets all the metadata is - # being pushed into the root. FIXME: MUST REMOVE SOON - key = prefix + key + "_" - for k in root_values.keys(): - if k.startswith( key ): - del root_values[k] - elif isinstance( input, Repeat ): - group_values = values[key] - for i, rep_values in enumerate( group_values ): - rep_index = rep_values['__index__'] - prefix = "%s_%d|" % ( key, rep_index ) - cleanup( prefix, input.inputs, group_values[i] ) - elif isinstance( input, Conditional ): - group_values = values[input.name] - current_case = group_values['__current_case__'] - prefix = "%s|" % ( key ) - cleanup( prefix, input.cases[current_case].inputs, group_values ) - cleanup( "", inputs, values ) - return associations - - - - -

1 0

[hg] galaxy 3073: Added UCI usage record and some automated page...
by Greg Von Kuster 23 Nov '09

23 Nov '09

details: http://www.bx.psu.edu/hg/galaxy/rev/6632a5d39f41 changeset: 3073:6632a5d39f41 user: Enis Afgan <afgane(a)gmail.com> date: Mon Oct 26 17:04:47 2009 -0400 description: Added UCI usage record and some automated page refreshes to cope with proper status reporting/visible links. diffstat: lib/galaxy/web/controllers/cloud.py | 11 +++- templates/cloud/configure_cloud.mako | 39 +++++++++++- templates/cloud/view.mako | 35 +++++++---- templates/cloud/view_usage.mako | 88 ++++++++++++++++++++++++++--- 4 files changed, 144 insertions(+), 29 deletions(-) diffs (285 lines): diff -r 676fae1a35f9 -r 6632a5d39f41 lib/galaxy/web/controllers/cloud.py --- a/lib/galaxy/web/controllers/cloud.py Fri Oct 23 17:41:35 2009 -0400 +++ b/lib/galaxy/web/controllers/cloud.py Mon Oct 26 17:04:47 2009 -0400 @@ -165,6 +165,10 @@ ( uci.state != uci_states.PENDING ) and \ ( uci.state != uci_states.DELETING ) and \ ( uci.state != uci_states.DELETING_UCI ) and \ + ( uci.state != uci_states.DELETED ) and \ + ( uci.state != uci_states.RUNNING ) and \ + ( uci.state != uci_states.NEW_UCI ) and \ + ( uci.state != uci_states.NEW ) and \ ( uci.state != uci_states.ERROR ): instance = model.CloudInstance() instance.user = user @@ -247,13 +251,16 @@ @web.expose @web.require_login( "use Galaxy cloud" ) - def usageReport( self, trans ): + def usageReport( self, trans, id ): user = trans.get_user() + id = trans.security.decode_id( id ) prevInstances = trans.sa_session.query( model.CloudInstance ) \ - .filter_by( user=user, state=instance_states.TERMINATED ) \ + .filter_by( user=user, state=instance_states.TERMINATED, uci_id=id ) \ .order_by( desc( model.CloudInstance.c.update_time ) ) \ .all() + + log.debug( "id: %s" % id ) return trans.fill_template( "cloud/view_usage.mako", prevInstances = prevInstances ) diff -r 676fae1a35f9 -r 6632a5d39f41 templates/cloud/configure_cloud.mako --- a/templates/cloud/configure_cloud.mako Fri Oct 23 17:41:35 2009 -0400 +++ b/templates/cloud/configure_cloud.mako Mon Oct 26 17:04:47 2009 -0400 @@ -24,16 +24,46 @@ $.getJSON( "${h.url_for( action='json_update' )}", {}, function ( data ) { for (var i in data) { var elem = '#' + data[i].id; + // Because of different list managing 'live' vs. 'available' instances, refresh entire + // page on necessary state change. + old_state = $(elem + "-state").text(); + new_state = data[i].state; + console.log( "old_state[%d] = %s", i, old_state ); + console.log( "new_state[%d] = %s", i, new_state ); + if ( old_state=='pending' && new_state=='running' ) { + location.reload(true); + } + else if ( old_state=='shutting-down' && new_state=='available' ) { + location.reload(true); + } + else if ( new_state=='shutting-down' || new_state=='shutting-downUCI' ) { + $(elem + "-link").text( "" ); + } + xmlhttp = new XMLHttpRequest(); + xmlhttp.open( "HEAD", "http://127.0.0.1:8080/admin", false ); + xmlhttp.send( null ); + //alert(xmlhttp.getAllResponseHeaders()) + console.log( "xmlhttp.readyState: %s", xmlhttp.readyState ); + console.log( "xmlhttp.status: %s", xmlhttp.status ); + if ( new_state=='running' && xmlhttp.readyState==1 ) { + console.log ("in ready statsus = 1"); + //if (xmlhttp.status==200) { + // console.log( "in status = 200" ); + // location.reload(true); + //} + } + + // Update 'state' and 'time alive' fields $(elem + "-state").text( data[i].state ); if (data[i].launch_time) { - $(elem + "-launch_time").text( data[i].launch_time.substring(0, 16 ) + " (" + data[i].time_ago + ")" ); + $(elem + "-launch_time").text( data[i].launch_time.substring(0, 16 ) + " UTC (" + data[i].time_ago + ")" ); } else { $(elem + "-launch_time").text( "N/A" ); } } }); - setTimeout("update_state()", 10000); + setTimeout("update_state()", 15000); } $(function() { @@ -141,7 +171,7 @@ context.write( ')' ) %> </td> - <td><div align="right"> + <td id="${ liveInstance.id }-link"><div align="right"> %for j, instance in enumerate( liveInstance.instance ): ## TODO: Once more instances will be running under the same liveInstance, additional logic will need to be added to account for that %if instance.state == "running": @@ -165,6 +195,7 @@ <a class="action-button" confirm="Are you sure you want to stop instance '${liveInstance.name}'? Please note that this may take up to 1 minute during which time the page will not refresh." href="${h.url_for( action='stop', id=trans.security.encode_id(liveInstance.id) )}">Stop</a> <a class="action-button" href="${h.url_for( action='renameInstance', id=trans.security.encode_id(liveInstance.id) )}">Rename</a> <a class="action-button" href="${h.url_for( action='viewInstance', id=trans.security.encode_id(liveInstance.id) )}">View details</a> + <a class="action-button" href="${h.url_for( action='usageReport', id=trans.security.encode_id(liveInstance.id) )}">Usage report</a> </div> </td> </tr> @@ -230,7 +261,7 @@ <a class="action-button" href="${h.url_for( action='start', id=trans.security.encode_id(prevInstance.id), type='c1.medium' )}"> Start c1.medium</a> <a class="action-button" href="${h.url_for( action='renameInstance', id=trans.security.encode_id(prevInstance.id) )}">Rename</a> <a class="action-button" href="${h.url_for( action='addStorage', id=trans.security.encode_id(prevInstance.id) )}" target="_parent">Add storage</a> - <a class="action-button" href="${h.url_for( action='usageReport' )}">Usage report</a> + <a class="action-button" href="${h.url_for( action='usageReport', id=trans.security.encode_id(prevInstance.id) )}">Usage report</a> <a class="action-button" confirm="Are you sure you want to delete instance '${prevInstance.name}'? This will delete all of your data assocaiated with this instance!" href="${h.url_for( action='deleteInstance', id=trans.security.encode_id(prevInstance.id) )}">Delete</a> </div> </td> diff -r 676fae1a35f9 -r 6632a5d39f41 templates/cloud/view.mako --- a/templates/cloud/view.mako Fri Oct 23 17:41:35 2009 -0400 +++ b/templates/cloud/view.mako Mon Oct 26 17:04:47 2009 -0400 @@ -1,20 +1,29 @@ <%inherit file="/base.mako"/> -<%def name="title()">Cloud home</%def> +<%def name="title()">Cloud credentials</%def> <h2>Credentials details</h2> +<ul class="manage-table-actions"> + <li> + <a class="action-button" href="${h.url_for( action='list' )}"> + <img src="${h.url_for('/static/images/silk/resultset_previous.png')}" /> + <span>Return to cloud management console</span> + </a> + </li> +</ul> + %if credDetails: - <ul class="manage-table-actions"> - <li> - <a class="action-button" href="${h.url_for( action='list' )}"> - <img src="${h.url_for('/static/images/silk/resultset_previous.png')}" /> - <span>Return to cloud management console</span> - </a> - </li> - </ul> + ${view_cred( credDetails )} +%else: + There are no credentials under that name. +%endif - <table class="mange-table colored" border="0" cellspacing="0" cellpadding="0" width="100%"> + + + +<%def name="view_cred( credDetails )"> + <table class="mange-table colored" border="0" cellspacing="0" cellpadding="0" width="100%"> <tr> <td> Credentials name: </td> <td> @@ -71,7 +80,7 @@ </div> </td> </tr> - <tr><td><b>Additional cloud provider information (if available):</b></td></tr> + <tr><td id="addl"><b>Additional cloud provider information (if available):</b></td></tr> %if credDetails.provider.region_connection != None: <tr> <td> Region connection: </td> @@ -145,6 +154,4 @@ </tr> %endif </table> -%else: - There are no credentials under that name. -%endif +</%def> diff -r 676fae1a35f9 -r 6632a5d39f41 templates/cloud/view_usage.mako --- a/templates/cloud/view_usage.mako Fri Oct 23 17:41:35 2009 -0400 +++ b/templates/cloud/view_usage.mako Mon Oct 26 17:04:47 2009 -0400 @@ -18,30 +18,100 @@ </div> %endif -<h2>Instance usage report</h2> +%if prevInstances: + <h2>Usage report for instance ${prevInstances[0].uci.name}</h2> +%else: + <h2>Selected instance has no record of being used.</h2> +%endif + +<ul class="manage-table-actions"> + <li> + <a class="action-button" href="${h.url_for( action='list' )}"> + <img src="${h.url_for('/static/images/silk/resultset_previous.png')}" /> + <span>Return to cloud management console</span> + </a> + </li> +</ul> %if prevInstances: <table class="mange-table colored" border="0" cellspacing="0" cellpadding="0" width="100%"> <colgroup width="2%"></colgroup> - <colgroup width="20%"></colgroup> - <colgroup width="20%"></colgroup> + <colgroup width="16%"></colgroup> + <colgroup width="16%"></colgroup> + <colgroup width="10%"></colgroup> <colgroup width="5%"></colgroup> <tr class="header"> <th>#</th> <th>Launch time</th> <th>Termination time</th> + <th>Time alive</th> <th>Type</th> <th></th> </tr> + <% + total_hours = 0 + %> %for i, prevInstance in enumerate( prevInstances ): <tr> <td>${i+1}</td> - <td>${str(prevInstance.launch_time)[:16]} UCT</td> - <td>${str(prevInstance.stop_time)[:16]} UCT</td> + <td> + %if prevInstance.launch_time: + ${str(prevInstance.launch_time)[:16]} UCT + %else: + N/A + %endif + </td> + <td> + %if prevInstance.stop_time: + ${str(prevInstance.stop_time)[:16]} UCT + %else: + N/A + %endif + </td> + <td> + <% + # This is where current time and since duration is calculated + if prevInstance.launch_time is None or prevInstance.stop_time is None: + context.write( 'N/A' ) + else: + context.write( str(h.date.distance_of_time_in_words (prevInstance.launch_time, prevInstance.stop_time ) ) ) + time_delta = prevInstance.stop_time - prevInstance.launch_time + total_hours += time_delta.seconds / 3600 + if time_delta.seconds != 0: + total_hours += 1 + + %> + </td> <td>${prevInstance.type}</td> - </tr> + </tr> %endfor </table> -%else: - Selected instance has no record of being used. -%endif \ No newline at end of file + <br/>Total number of hours instance was alive: ${total_hours} <br /> + Note that these are just best effort estimates - true usage should be obtained from respective cloud provider. <br /> + <%namespace name="view_cred" file="view.mako" /> + + <div id="hide_cred_details"> + This instance uses credentials: + <a onclick="document.getElementById('show_cred_details').style.display = 'block'; + document.getElementById('hide_cred_details').style.display = 'none'; return 0" + href="javascript:void(0)"> + ${prevInstances[0].uci.credentials.name} + </a> + </div> + <div id="show_cred_details" style="DISPLAY: none"> + This instance uses credentials: + <a onclick="document.getElementById('hide_cred_details').style.display = 'block'; + document.getElementById('show_cred_details').style.display = 'none'; return 0;" + href="javascript:void(0)"> + ${prevInstances[0].uci.credentials.name} + </a> + ${view_cred.view_cred( prevInstances[0].uci.credentials ) } + </div> + + +%endif + + + + +

1 0

[hg] galaxy 3075: Added editing of credential properties.
by Greg Von Kuster 23 Nov '09

23 Nov '09

details: http://www.bx.psu.edu/hg/galaxy/rev/7b642483d18c changeset: 3075:7b642483d18c user: Enis Afgan <afgane(a)gmail.com> date: Wed Oct 28 12:28:57 2009 -0400 description: Added editing of credential properties. diffstat: lib/galaxy/web/controllers/cloud.py | 91 +++++++++++++++++++++++------ templates/cloud/add_credentials.mako | 40 ++++++------- templates/cloud/configure_cloud.mako | 2 +- templates/cloud/edit_credentials.mako | 91 ++++++++++++++++++++++++++++++ templates/cloud/view.mako | 6 +- 5 files changed, 185 insertions(+), 45 deletions(-) diffs (315 lines): diff -r 802b761c5032 -r 7b642483d18c lib/galaxy/web/controllers/cloud.py --- a/lib/galaxy/web/controllers/cloud.py Tue Oct 27 15:32:44 2009 -0400 +++ b/lib/galaxy/web/controllers/cloud.py Wed Oct 28 12:28:57 2009 -0400 @@ -77,7 +77,7 @@ cloudCredentials = trans.sa_session.query( model.CloudUserCredentials ) \ .filter_by( user=user ) \ - .order_by( desc( model.CloudUserCredentials.c.name ) ) \ + .order_by( model.CloudUserCredentials.c.name ) \ .all() liveInstances = trans.sa_session.query( model.UCI ) \ @@ -387,23 +387,64 @@ @web.expose @web.require_login( "use Galaxy cloud" ) - def rename( self, trans, id, new_name=None ): - stored = get_stored_credentials( trans, id ) - if new_name is not None: - stored.name = new_name - trans.sa_session.flush() - trans.set_message( "Credentials renamed to '%s'." % new_name ) - return self.list( trans ) + def edit( self, trans, id, credName=None, accessKey=None, secretKey=None, edited=False ): + error = {} + if not edited: + credentials = get_stored_credentials( trans, id ) + return trans.fill_template( "cloud/edit_credentials.mako", + credential = credentials, + error = error + ) else: - return trans.show_form( - web.FormBuilder( url_for( id=trans.security.encode_id(stored.id) ), "Rename credentials", submit_text="Rename" ) - .add_text( "new_name", "Credentials Name", value=stored.name ) ) - + user = trans.get_user() + credentials = get_stored_credentials( trans, id ) + if credName=='' or len( credName ) > 255: + error['cred_error'] = "Credentials name must be between 1 and 255 characters in length." + elif trans.app.model.CloudUserCredentials \ + .filter_by( user=user ) \ + .filter( and_( trans.app.model.CloudUserCredentials.table.c.id != credentials.id, trans.app.model.CloudUserCredentials.table.c.name==credName ) ) \ + .first(): + error['cred_error'] = "Credentials with name '" + credName + "' already exist. Please choose an alternative name." + elif accessKey=='' or len( accessKey ) > 255: + error['access_key_error'] = "Access key must be between 1 and 255 characters long." + elif secretKey=='' or len( secretKey ) > 255: + error['secret_key_error'] = "Secret key must be between 1 and 255 characters long." + + if error: + return trans.fill_template( "cloud/edit_credentials.mako", + credential = credentials, + error = error + ) + else: + # Edit user stored credentials + credentials.name = credName + credentials.access_key = accessKey + credentials.secret_key = secretKey + # Persist + session = trans.sa_session + session.save_or_update( credentials ) + session.flush() + # Log and display the management page + trans.set_message( "Credential '%s' edited." % credentials.name ) +# if defaultCred: +# self.makeDefault( trans, credentials.id) + return self.list( trans ) + @web.expose @web.require_login( "use Galaxy cloud" ) def renameInstance( self, trans, id, new_name=None ): instance = get_uci( trans, id ) if new_name is not None: + if len(new_name) > 255: + error( "Instance name must be less than 255 characters long." ) + user = trans.get_user() + name_exists = trans.sa_session.query( model.UCI ) \ + .filter_by( user=user, name=new_name ) \ + .first() + if name_exists: + error( "Specified name ('%s') is already used by an existing instance. Please choose an alternative name." % new_name ) + + # Update name in local DB instance.name = new_name trans.sa_session.flush() trans.set_message( "Instance renamed to '%s'." % new_name ) @@ -522,17 +563,27 @@ @web.require_login( "delete credentials" ) def delete( self, trans, id=None ): """ - Delete user's cloud credentials - TODO: Because UCI's depend on specific credentials, need to handle case where given credentials are being used by a UCI + Delete user's cloud credentials checking that no registered instances are tied to given credentials. """ # Load credentials from database + user = trans.get_user() stored = get_stored_credentials( trans, id ) - # Delete and save - sess = trans.sa_session - sess.delete( stored ) - stored.flush() - # Display the management page - trans.set_message( "Credentials '%s' deleted." % stored.name ) + UCIs = trans.sa_session.query( model.UCI ) \ + .filter_by( user=user, credentials_id=stored.id ) \ + .filter( model.UCI.c.state!=uci_states.DELETED ) \ + .all() + + if len(UCIs) == 0: + # Delete and save + sess = trans.sa_session + sess.delete( stored ) + stored.flush() + # Display the management page + trans.set_message( "Credentials '%s' deleted." % stored.name ) + return self.list( trans ) + + error( "Existing instance(s) depend on credentials '%s'. You must delete those instances before being able \ + to delete these credentials." % stored.name ) return self.list( trans ) @web.expose diff -r 802b761c5032 -r 7b642483d18c templates/cloud/add_credentials.mako --- a/templates/cloud/add_credentials.mako Tue Oct 27 15:32:44 2009 -0400 +++ b/templates/cloud/add_credentials.mako Wed Oct 28 12:28:57 2009 -0400 @@ -20,8 +20,8 @@ <div class="form"> <div class="form-title">Add credentials</div> <div class="form-body"> - <form name="Add credentials" action="${h.url_for( action='add' )}" method="post" > - + <form name="Add credentials" action="${h.url_for( action='add' )}" method="post" > + <% cls = "form-row" if error.has_key('cred_error'): @@ -44,23 +44,23 @@ if error.has_key('provider_error'): cls += " form-row-error" %> - <div class="${cls}"> - <label>Cloud provider name:</label> - <div class="form-row-input"> - <select name="providerName" style="width:40em"> - <option value="">Select Provider...</option> - %for provider in providers: - <option value="${provider.name}">${provider.name}</option> - %endfor - </select> - <br/>or <a href="${h.url_for( action='add_provider' )}"> - <span>register additional cloud provider</span></a> - </div> - %if error.has_key('provider_error'): - <div class="form-row-error-message">${error['provider_error']}</div> - %endif - <div style="clear: both"></div> - </div> + <div class="${cls}"> + <label>Cloud provider name:</label> + <div class="form-row-input"> + <select name="providerName" style="width:40em"> + <option value="">Select Provider...</option> + %for provider in providers: + <option value="${provider.name}">${provider.name}</option> + %endfor + </select> + <br/>or <a href="${h.url_for( action='add_provider' )}"> + <span>register additional cloud provider</span></a> + </div> + %if error.has_key('provider_error'): + <div class="form-row-error-message">${error['provider_error']}</div> + %endif + <div style="clear: both"></div> + </div> <% cls = "form-row" @@ -95,9 +95,7 @@ <div style="clear: both"></div> </div> - <div class="form-row"><input type="submit" value="Add"></div> - </form> </div> </div> diff -r 802b761c5032 -r 7b642483d18c templates/cloud/configure_cloud.mako --- a/templates/cloud/configure_cloud.mako Tue Oct 27 15:32:44 2009 -0400 +++ b/templates/cloud/configure_cloud.mako Wed Oct 28 12:28:57 2009 -0400 @@ -109,7 +109,7 @@ <td> <div popupmenu="cr-${i}-popup"> <a class="action-button" href="${h.url_for( action='view', id=trans.security.encode_id(cloudCredential.id) )}">View</a> - <a class="action-button" href="${h.url_for( action='rename', id=trans.security.encode_id(cloudCredential.id) )}">Rename</a> + <a class="action-button" href="${h.url_for( action='edit', id=trans.security.encode_id(cloudCredential.id) )}">Edit</a> <a class="action-button" confirm="Are you sure you want to delete credentials '${cloudCredential.name}'?" href="${h.url_for( action='delete', id=trans.security.encode_id(cloudCredential.id) )}">Delete</a> </div> </td> diff -r 802b761c5032 -r 7b642483d18c templates/cloud/edit_credentials.mako --- /dev/null Thu Jan 01 00:00:00 1970 +0000 +++ b/templates/cloud/edit_credentials.mako Wed Oct 28 12:28:57 2009 -0400 @@ -0,0 +1,91 @@ +<% _=n_ %> +<%inherit file="/base.mako"/> +<%def name="title()">Add credentials</%def> + +<%def name="javascripts()"> +${parent.javascripts()} +<script type="text/javascript"> +$(function(){ + $("input:text:first").focus(); +}) +</script> +</%def> + +%if header: + ${header} +%endif + +%if credential: + +<div class="form"> + <div class="form-title">Edit credentials</div> + <div class="form-body"> + <form name="edit_credentials" action="${h.url_for( action='edit', id=trans.security.encode_id(credential.id), edited="true" )}" method="post" > + + <% + cls = "form-row" + if error.has_key('cred_error'): + cls += " form-row-error" + %> + <div class="${cls}"> + <label>Credentials name:</label> + <div class="form-row-input"> + <input type="text" name="credName" value="${credential.name}" size="40"> + </div> + %if error.has_key('cred_error'): + <div class="form-row-error-message">${error['cred_error']}</div> + %endif + <div style="clear: both"></div> + </div> + + + <% + cls = "form-row" + if error.has_key('provider_error'): + cls += " form-row-error" + %> + <div class="${cls}"> + <label>Cloud provider name (type):</label> + <div class="form-row-input">${credential.provider.name} (${credential.provider.type})</div> + <div style="clear: both"></div> + </div> + <% + cls = "form-row" + if error.has_key('access_key_error'): + cls += " form-row-error" + %> + <div class="${cls}"> + <label>Access key:</label> + <div class="form-row-input"> + <input type="text" name="accessKey" value="${credential.access_key}" size="40"> + </div> + %if error.has_key('access_key_error'): + <div class="form-row-error-message">${error['access_key_error']}</div> + %endif + <div style="clear: both"></div> + </div> + + <% + cls = "form-row" + if error.has_key('secret_key_error'): + cls += " form-row-error" + %> + <div class="${cls}"> + <label>Secret key:</label> + <div class="form-row-input"> + <input type="password" name="secretKey" value="${credential.secret_key}" size="40"> + </div> + %if error.has_key('secret_key_error'): + <div class="form-row-error-message">${error['secret_key_error']}</div> + %endif + <div style="clear: both"></div> + </div> + + <div class="form-row"><input type="submit" value="Edit"></div> + </form> + </div> +</div> + +%else: + Specified credentials could not be found. +%endif diff -r 802b761c5032 -r 7b642483d18c templates/cloud/view.mako --- a/templates/cloud/view.mako Tue Oct 27 15:32:44 2009 -0400 +++ b/templates/cloud/view.mako Wed Oct 28 12:28:57 2009 -0400 @@ -71,12 +71,12 @@ </a> </div> <div id="fullComment2" style="DISPLAY: none"> - <nobr><b>${credDetails.secret_key}</b></nobr><br/> - <a onclick="document.getElementById('shortComment2').style.display = 'block'; + <a onclick="document.getElementById('shortComment2').style.display = 'block'; document.getElementById('fullComment2').style.display = 'none'; return 0;" href="javascript:void(0)"> - Hide - </a> + </a><br /> + <nobr><b>${credDetails.secret_key}</b></nobr><br/> </div> </td> </tr>

1 0

[hg] galaxy 3074: UI enhancements regarding provider registratio...
by Greg Von Kuster 23 Nov '09

23 Nov '09

details: http://www.bx.psu.edu/hg/galaxy/rev/802b761c5032 changeset: 3074:802b761c5032 user: Enis Afgan <afgane(a)gmail.com> date: Tue Oct 27 15:32:44 2009 -0400 description: UI enhancements regarding provider registration and error reporting. diffstat: lib/galaxy/cloud/providers/ec2.py | 37 +- lib/galaxy/cloud/providers/eucalyptus.py | 15 +- lib/galaxy/web/controllers/cloud.py | 553 ++++++------------------------------ templates/cloud/add_credentials.mako | 4 +- templates/cloud/add_provider.mako | 52 ++- templates/cloud/configure_cloud.mako | 43 +- 6 files changed, 200 insertions(+), 504 deletions(-) diffs (922 lines): diff -r 6632a5d39f41 -r 802b761c5032 lib/galaxy/cloud/providers/ec2.py --- a/lib/galaxy/cloud/providers/ec2.py Mon Oct 26 17:04:47 2009 -0400 +++ b/lib/galaxy/cloud/providers/ec2.py Tue Oct 27 15:32:44 2009 -0400 @@ -95,7 +95,13 @@ Establishes EC2 cloud connection using user's credentials associated with given UCI """ log.debug( '##### Establishing EC2 cloud connection' ) - conn = EC2Connection( uci_wrapper.get_access_key(), uci_wrapper.get_secret_key() ) + provider = uci_wrapper.get_provider() + region = RegionInfo( None, provider.region_name, provider.region_endpoint ) + conn = EC2Connection( aws_access_key_id=uci_wrapper.get_access_key(), + aws_secret_access_key=uci_wrapper.get_secret_key(), + is_secure=provider.is_secure, + region=region, + path=provider.path ) return conn def set_keypair( self, uci_wrapper, conn ): @@ -458,16 +464,24 @@ a_key = uci.credentials.access_key s_key = uci.credentials.secret_key # Get connection - conn = EC2Connection( aws_access_key_id=a_key, aws_secret_access_key=s_key ) + region = RegionInfo( None, uci.credentials.provider.region_name, uci.credentials.provider.region_endpoint ) + conn = EC2Connection( aws_access_key_id=a_key, + aws_secret_access_key=s_key, + is_secure=uci.credentials.provider.is_secure, + region=region, + path=uci.credentials.provider.path ) # Get reservations handle for given instance rl= conn.get_all_instances( [inst.instance_id] ) # Because EPC deletes references to reservations after a short while after instances have terminated, getting an empty list as a response to a query - # typically means the instance has successfully shut down but the check was not performed in short enough amount of time. As a result, below code simply - # marks given instance as having terminated. Note that an instance might have also crashed and this code will not catch the difference... + # typically means the instance has successfully shut down but the check was not performed in short enough amount of time. Until alternative solution + # is found, below code sets state of given UCI to 'error' to indicate to the user something out of ordinary happened. if len( rl ) == 0: log.info( "Instance ID '%s' was not found by the cloud provider. Instance might have crashed or otherwise been terminated." % inst.instance_id ) + inst.error = "Instance ID was not found by the cloud provider. Instance might have crashed or otherwise been terminated. State set to 'terminated'." + uci.error = "Instance ID '"+inst.instance_id+"' was not found by the cloud provider. Instance might have crashed or otherwise been terminated."+ \ + "Manual check is recommended." inst.state = instance_states.TERMINATED - uci.state = uci_states.AVAILABLE + uci.state = uci_states.ERROR uci.launch_time = None inst.flush() uci.flush() @@ -479,11 +493,13 @@ if s != inst.state: inst.state = s inst.flush() - if s == instance_states.TERMINATED: # After instance has shut down, ensure UCI is marked as 'available' + # After instance has shut down, ensure UCI is marked as 'available' + if s == instance_states.TERMINATED and uci.state != uci_states.ERROR: uci.state = uci_states.AVAILABLE + uci.launch_time = None uci.flush() + # Making sure state of UCI is updated. Once multiple instances become associated with single UCI, this will need to be changed. if s != uci.state and s != instance_states.TERMINATED: - # Making sure state of UCI is updated. Once multiple instances become associated with single UCI, this will need to be changed. uci.state = s uci.flush() if cInst.public_dns_name != inst.public_dns: @@ -501,7 +517,12 @@ a_key = uci.credentials.access_key s_key = uci.credentials.secret_key # Get connection - conn = EC2Connection( aws_access_key_id=a_key, aws_secret_access_key=s_key ) + region = RegionInfo( None, uci.credentials.provider.region_name, uci.credentials.provider.region_endpoint ) + conn = EC2Connection( aws_access_key_id=a_key, + aws_secret_access_key=s_key, + is_secure=uci.credentials.provider.is_secure, + region=region, + path=uci.credentials.provider.path ) # Get reservations handle for given store vl = conn.get_all_volumes( [store.volume_id] ) # log.debug( "Store '%s' vl: '%s'" % ( store.volume_id, vl ) ) diff -r 6632a5d39f41 -r 802b761c5032 lib/galaxy/cloud/providers/eucalyptus.py --- a/lib/galaxy/cloud/providers/eucalyptus.py Mon Oct 26 17:04:47 2009 -0400 +++ b/lib/galaxy/cloud/providers/eucalyptus.py Tue Oct 27 15:32:44 2009 -0400 @@ -443,12 +443,15 @@ # Get reservations handle for given instance rl= conn.get_all_instances( [inst.instance_id] ) # Because EPC deletes references to reservations after a short while after instances have terminated, getting an empty list as a response to a query - # typically means the instance has successfully shut down but the check was not performed in short enough amount of time. As a result, below code simply - # marks given instance as having terminated. Note that an instance might have also crashed and this code will not catch the difference... + # typically means the instance has successfully shut down but the check was not performed in short enough amount of time. Until alternative solution + # is found, below code sets state of given UCI to 'error' to indicate to the user something out of ordinary happened. if len( rl ) == 0: log.info( "Instance ID '%s' was not found by the cloud provider. Instance might have crashed or otherwise been terminated." % inst.instance_id ) + inst.error = "Instance ID was not found by the cloud provider. Instance might have crashed or otherwise been terminated. State set to 'terminated'." + uci.error = "Instance ID '"+inst.instance_id+"' was not found by the cloud provider. Instance might have crashed or otherwise been terminated."+ \ + "Manual check is recommended." inst.state = instance_states.TERMINATED - uci.state = uci_states.AVAILABLE + uci.state = uci_states.ERROR uci.launch_time = None inst.flush() uci.flush() @@ -460,11 +463,13 @@ if s != inst.state: inst.state = s inst.flush() - if s == instance_states.TERMINATED: # After instance has shut down, ensure UCI is marked as 'available' + # After instance has shut down, ensure UCI is marked as 'available' + if s == instance_states.TERMINATED and uci.state != uci_states.ERROR: uci.state = uci_states.AVAILABLE + uci.launch_time = None uci.flush() + # Making sure state of UCI is updated. Once multiple instances become associated with single UCI, this will need to be changed. if s != uci.state and s != instance_states.TERMINATED: - # Making sure state of UCI is updated. Once multiple instances become associated with single UCI, this will need to be changed. uci.state = s uci.flush() if cInst.public_dns_name != inst.public_dns: diff -r 6632a5d39f41 -r 802b761c5032 lib/galaxy/web/controllers/cloud.py --- a/lib/galaxy/web/controllers/cloud.py Mon Oct 26 17:04:47 2009 -0400 +++ b/lib/galaxy/web/controllers/cloud.py Tue Oct 27 15:32:44 2009 -0400 @@ -273,11 +273,10 @@ inst_error = vol_error = cred_error = None error = {} user = trans.get_user() - # TODO: Hack until present user w/ bullet list w/ registered credentials storedCreds = trans.sa_session.query( model.CloudUserCredentials ).filter_by( user=user ).all() if len( storedCreds ) == 0: return trans.show_error_message( "You must register credentials before configuring a Galaxy instance." ) - + # Create dict mapping of cloud providers to zones available by those providers providersToZones = {} for storedCred in storedCreds: if storedCred.provider.type == 'ec2': @@ -413,6 +412,21 @@ return trans.show_form( web.FormBuilder( url_for( id=trans.security.encode_id(instance.id) ), "Rename instance", submit_text="Rename" ) .add_text( "new_name", "Instance name", value=instance.name ) ) + + @web.expose + @web.require_login( "use Galaxy cloud" ) + def set_uci_state( self, trans, id, state='available', clear_error=True ): + """ + Sets state of UCI to given state, optionally resets error field, and resets UCI's launch time field to 'None'. + """ + uci = get_uci( trans, id ) + uci.state = state + if clear_error: + uci.error = None + uci.launch_time = None + trans.sa_session.flush() + trans.set_message( "Instance '%s' state reset." % uci.name ) + return self.list( trans ) @web.expose @web.require_login( "add credentials" ) @@ -527,42 +541,94 @@ proxy_user='', proxy_pass='', debug='', https_connection_factory='', path='' ): user = trans.get_user() error = {} + try: + is_secure = int(is_secure) + except ValueError: + pass # Check if Amazon EC2 has already been registered by this user ec2_registered = trans.sa_session.query( model.CloudProvider ).filter_by( user=user, type='ec2' ).first() if region_name or region_endpoint or name or is_secure or port or proxy or debug or path: + log.debug (" in if ") if trans.app.model.CloudProvider \ .filter_by (user=user, name=name) \ .first(): + log.debug (" in if 2 ") error['name_error'] = "A provider with that name already exist." elif name=='' or len( name ) > 255: + log.debug (" in if 3") error['name_error'] = "Provider name must be between 1 and 255 characters long." elif type=='': + log.debug (" in if 4") error['type_error'] = "Provider type must be selected." elif ec2_registered: + log.debug (" in if 5") error['type_error'] = "Amazon EC2 has already been registered as a provider." - elif is_secure!=0 or is_secure!=1: - error['is_secure_error'] = "Field 'is secure' can only take on a value '0' or '1'." + elif not (is_secure == 0 or is_secure == 1): + log.debug (" in if 6") + error['is_secure_error'] = "Field 'is secure' can only take on a value '0' or '1'" else: + log.debug (" in else ") provider = model.CloudProvider() provider.user = user provider.type = type provider.name = name - provider.region_name = region_name - provider.region_endpoint = region_endpoint + if region_name: + provider.region_name = region_name + else: + provider.region_name = None + + if region_endpoint: + provider.region_endpoint = region_endpoint + else: + provider.region_endpoint = None + if is_secure=='0': provider.is_secure = False else: provider.is_secure = True - provider.host = host - provider.port = port - provider.proxy = proxy - provider.proxy_port = proxy_port - provider.proxy_user = proxy_user - provider.proxy_pass = proxy_pass - provider.debug = debug - provider.https_connection_factory = https_connection_factory + + if host: + provider.host = host + else: + provider.host = None + + if port: + provider.port = port + else: + provider.port = None + + if proxy: + provider.proxy = proxy + else: + provider.proxy = None + + if proxy_port: + provider.proxy_port = proxy_port + else: + provider.proxy_port = None + + if proxy_user: + provider.proxy_user = proxy_user + else: + provider.proxy_user = None + + if proxy_pass: + provider.proxy_pass = proxy_pass + else: + provider.proxy_pass = None + + if debug: + provider.debug = debug + else: + provider.debug = None + + if https_connection_factory: + provider.https_connection_factory = https_connection_factory + else: + provider.https_connection_factory = None + provider.path = path # Persist session = trans.sa_session @@ -601,456 +667,21 @@ .filter_by( user=user, type='ec2' ).first() if not exists: - provider = model.CloudProvider() - provider.user = user - provider.type = 'ec2' - provider.name = 'EC2' - # Persist - session = trans.sa_session - session.save_or_update( provider ) - session.flush() + self.add_provider( trans, name='Amazon EC2', type='ec2', region_name='us-east-1', region_endpoint='us-east-1.ec2.amazonaws.com', is_secure=1, path='/' ) + return self.add( trans ) +# providers = trans.sa_session.query( model.CloudProvider ).filter_by( user=user ).all() +# return trans.fill_template( "cloud/add_credentials.mako", +# credName = '', +# providerName = '', +# accessKey = '', +# secretKey = '', +# error = {}, +# providers = providers +# ) + trans.show_error_message( "EC2 is already registered as a cloud provider under name '%s'." % exists.name ) return self.list( trans ) - @web.expose - @web.require_login( "edit workflows" ) - def editor( self, trans, id=None ): - """ - Render the main workflow editor interface. The canvas is embedded as - an iframe (neccesary for scrolling to work properly), which is - rendered by `editor_canvas`. - """ - if not id: - error( "Invalid workflow id" ) - id = trans.security.decode_id( id ) - return trans.fill_template( "workflow/editor.mako", workflow_id=id ) - - @web.json - def editor_form_post( self, trans, type='tool', tool_id=None, **incoming ): - """ - Accepts a tool state and incoming values, and generates a new tool - form and some additional information, packed into a json dictionary. - This is used for the form shown in the right pane when a node - is selected. - """ - trans.workflow_building_mode = True - module = module_factory.from_dict( trans, { - 'type': type, - 'tool_id': tool_id, - 'tool_state': incoming.pop("tool_state") - } ) - module.update_state( incoming ) - return { - 'tool_state': module.get_state(), - 'data_inputs': module.get_data_inputs(), - 'data_outputs': module.get_data_outputs(), - 'tool_errors': module.get_errors(), - 'form_html': module.get_config_form() - } - - @web.json - def get_new_module_info( self, trans, type, **kwargs ): - """ - Get the info for a new instance of a module initialized with default - paramters (any keyword arguments will be passed along to the module). - Result includes data inputs and outputs, html representation - of the initial form, and the initial tool state (with default values). - This is called asynchronously whenever a new node is added. - """ - trans.workflow_building_mode = True - module = module_factory.new( trans, type, **kwargs ) - return { - 'type': module.type, - 'name': module.get_name(), - 'tool_id': module.get_tool_id(), - 'tool_state': module.get_state(), - 'data_inputs': module.get_data_inputs(), - 'data_outputs': module.get_data_outputs(), - 'form_html': module.get_config_form() - } - - @web.json - def load_workflow( self, trans, id ): - """ - Get the latest Workflow for the StoredWorkflow identified by `id` and - encode it as a json string that can be read by the workflow editor - web interface. - """ - user = trans.get_user() - id = trans.security.decode_id( id ) - trans.workflow_building_mode = True - # Load encoded workflow from database - stored = trans.sa_session.query( model.StoredWorkflow ).get( id ) - assert stored.user == user - workflow = stored.latest_workflow - # Pack workflow data into a dictionary and return - data = {} - data['name'] = workflow.name - data['steps'] = {} - data['upgrade_messages'] = {} - # For each step, rebuild the form and encode the state - for step in workflow.steps: - # Load from database representation - module = module_factory.from_workflow_step( trans, step ) - # Fix any missing parameters - upgrade_message = module.check_and_update_state() - if upgrade_message: - data['upgrade_messages'][step.order_index] = upgrade_message - # Pack atrributes into plain dictionary - step_dict = { - 'id': step.order_index, - 'type': module.type, - 'tool_id': module.get_tool_id(), - 'name': module.get_name(), - 'tool_state': module.get_state(), - 'tool_errors': module.get_errors(), - 'data_inputs': module.get_data_inputs(), - 'data_outputs': module.get_data_outputs(), - 'form_html': module.get_config_form(), - } - # Connections - input_conn_dict = {} - for conn in step.input_connections: - input_conn_dict[ conn.input_name ] = \ - dict( id=conn.output_step.order_index, output_name=conn.output_name ) - step_dict['input_connections'] = input_conn_dict - # Position - step_dict['position'] = step.position - # Add to return value - data['steps'][step.order_index] = step_dict - print data['upgrade_messages'] - return data - - @web.json - def save_workflow( self, trans, id, workflow_data ): - """ - Save the workflow described by `workflow_data` with id `id`. - """ - # Get the stored workflow - stored = get_stored_workflow( trans, id ) - # Put parameters in workflow mode - trans.workflow_building_mode = True - # Convert incoming workflow data from json - data = simplejson.loads( workflow_data ) - # Create new workflow from incoming data - workflow = model.Workflow() - # Just keep the last name (user can rename later) - workflow.name = stored.name - # Assume no errors until we find a step that has some - workflow.has_errors = False - # Create each step - steps = [] - # The editor will provide ids for each step that we don't need to save, - # but do need to use to make connections - steps_by_external_id = {} - # First pass to build step objects and populate basic values - for key, step_dict in data['steps'].iteritems(): - # Create the model class for the step - step = model.WorkflowStep() - steps.append( step ) - steps_by_external_id[ step_dict['id' ] ] = step - # FIXME: Position should be handled inside module - step.position = step_dict['position'] - module = module_factory.from_dict( trans, step_dict ) - module.save_to_step( step ) - if step.tool_errors: - workflow.has_errors = True - # Stick this in the step temporarily - step.temp_input_connections = step_dict['input_connections'] - # Second pass to deal with connections between steps - for step in steps: - # Input connections - for input_name, conn_dict in step.temp_input_connections.iteritems(): - if conn_dict: - conn = model.WorkflowStepConnection() - conn.input_step = step - conn.input_name = input_name - conn.output_name = conn_dict['output_name'] - conn.output_step = steps_by_external_id[ conn_dict['id'] ] - del step.temp_input_connections - # Order the steps if possible - attach_ordered_steps( workflow, steps ) - # Connect up - workflow.stored_workflow = stored - stored.latest_workflow = workflow - # Persist - trans.sa_session.flush() - # Return something informative - errors = [] - if workflow.has_errors: - errors.append( "Some steps in this workflow have validation errors" ) - if workflow.has_cycles: - errors.append( "This workflow contains cycles" ) - if errors: - rval = dict( message="Workflow saved, but will not be runnable due to the following errors", - errors=errors ) - else: - rval = dict( message="Workflow saved" ) - rval['name'] = workflow.name - return rval - - @web.json - def get_datatypes( self, trans ): - ext_to_class_name = dict() - classes = [] - for k, v in trans.app.datatypes_registry.datatypes_by_extension.iteritems(): - c = v.__class__ - ext_to_class_name[k] = c.__module__ + "." + c.__name__ - classes.append( c ) - class_to_classes = dict() - def visit_bases( types, cls ): - for base in cls.__bases__: - if issubclass( base, Data ): - types.add( base.__module__ + "." + base.__name__ ) - visit_bases( types, base ) - for c in classes: - n = c.__module__ + "." + c.__name__ - types = set( [ n ] ) - visit_bases( types, c ) - class_to_classes[ n ] = dict( ( t, True ) for t in types ) - return dict( ext_to_class_name=ext_to_class_name, class_to_classes=class_to_classes ) - - @web.expose - def build_from_current_history( self, trans, job_ids=None, dataset_ids=None, workflow_name=None ): - user = trans.get_user() - history = trans.get_history() - if not user: - return trans.show_error_message( "Must be logged in to create workflows" ) - if ( job_ids is None and dataset_ids is None ) or workflow_name is None: - jobs, warnings = get_job_dict( trans ) - # Render - return trans.fill_template( - "workflow/build_from_current_history.mako", - jobs=jobs, - warnings=warnings, - history=history ) - else: - # Ensure job_ids and dataset_ids are lists (possibly empty) - if job_ids is None: - job_ids = [] - elif type( job_ids ) is not list: - job_ids = [ job_ids ] - if dataset_ids is None: - dataset_ids = [] - elif type( dataset_ids ) is not list: - dataset_ids = [ dataset_ids ] - # Convert both sets of ids to integers - job_ids = [ int( id ) for id in job_ids ] - dataset_ids = [ int( id ) for id in dataset_ids ] - # Find each job, for security we (implicately) check that they are - # associated witha job in the current history. - jobs, warnings = get_job_dict( trans ) - jobs_by_id = dict( ( job.id, job ) for job in jobs.keys() ) - steps = [] - steps_by_job_id = {} - hid_to_output_pair = {} - # Input dataset steps - for hid in dataset_ids: - step = model.WorkflowStep() - step.type = 'data_input' - hid_to_output_pair[ hid ] = ( step, 'output' ) - steps.append( step ) - # Tool steps - for job_id in job_ids: - assert job_id in jobs_by_id, "Attempt to create workflow with job not connected to current history" - job = jobs_by_id[ job_id ] - tool = trans.app.toolbox.tools_by_id[ job.tool_id ] - param_values = job.get_param_values( trans.app ) - associations = cleanup_param_values( tool.inputs, param_values ) - step = model.WorkflowStep() - step.type = 'tool' - step.tool_id = job.tool_id - step.tool_inputs = tool.params_to_strings( param_values, trans.app ) - # NOTE: We shouldn't need to do two passes here since only - # an earlier job can be used as an input to a later - # job. - for other_hid, input_name in associations: - if other_hid in hid_to_output_pair: - other_step, other_name = hid_to_output_pair[ other_hid ] - conn = model.WorkflowStepConnection() - conn.input_step = step - conn.input_name = input_name - # Should always be connected to an earlier step - conn.output_step = other_step - conn.output_name = other_name - steps.append( step ) - steps_by_job_id[ job_id ] = step - # Store created dataset hids - for assoc in job.output_datasets: - hid_to_output_pair[ assoc.dataset.hid ] = ( step, assoc.name ) - # Workflow to populate - workflow = model.Workflow() - workflow.name = workflow_name - # Order the steps if possible - attach_ordered_steps( workflow, steps ) - # And let's try to set up some reasonable locations on the canvas - # (these are pretty arbitrary values) - levorder = order_workflow_steps_with_levels( steps ) - base_pos = 10 - for i, steps_at_level in enumerate( levorder ): - for j, index in enumerate( steps_at_level ): - step = steps[ index ] - step.position = dict( top = ( base_pos + 120 * j ), - left = ( base_pos + 220 * i ) ) - # Store it - stored = model.StoredWorkflow() - stored.user = user - stored.name = workflow_name - workflow.stored_workflow = stored - stored.latest_workflow = workflow - trans.sa_session.save_or_update( stored ) - trans.sa_session.flush() - # Index page with message - return trans.show_message( "Workflow '%s' created from current history." % workflow_name ) - ## return trans.show_ok_message( "<p>Workflow '%s' created.</p><p><a target='_top' href='%s'>Click to load in workflow editor</a></p>" - ## % ( workflow_name, web.url_for( action='editor', id=trans.security.encode_id(stored.id) ) ) ) - - @web.expose - def run( self, trans, id, check_user=True, **kwargs ): - stored = get_stored_workflow( trans, id, check_ownership=False ) - if check_user: - user = trans.get_user() - if stored.user != user: - if trans.sa_session.query( model.StoredWorkflowUserShareAssociation ) \ - .filter_by( user=user, stored_workflow=stored ).count() == 0: - error( "Workflow is not owned by or shared with current user" ) - # Get the latest revision - workflow = stored.latest_workflow - # It is possible for a workflow to have 0 steps - if len( workflow.steps ) == 0: - error( "Workflow cannot be run because it does not have any steps" ) - #workflow = Workflow.from_simple( simplejson.loads( stored.encoded_value ), trans.app ) - if workflow.has_cycles: - error( "Workflow cannot be run because it contains cycles" ) - if workflow.has_errors: - error( "Workflow cannot be run because of validation errors in some steps" ) - # Build the state for each step - errors = {} - if kwargs: - # If kwargs were provided, the states for each step should have - # been POSTed - for step in workflow.steps: - # Connections by input name - step.input_connections_by_name = \ - dict( ( conn.input_name, conn ) for conn in step.input_connections ) - # Extract just the arguments for this step by prefix - p = "%s|" % step.id - l = len(p) - step_args = dict( ( k[l:], v ) for ( k, v ) in kwargs.iteritems() if k.startswith( p ) ) - step_errors = None - if step.type == 'tool' or step.type is None: - module = module_factory.from_workflow_step( trans, step ) - # Any connected input needs to have value DummyDataset (these - # are not persisted so we need to do it every time) - module.add_dummy_datasets( connections=step.input_connections ) - # Get the tool - tool = module.tool - # Get the state - step.state = state = module.state - # Get old errors - old_errors = state.inputs.pop( "__errors__", {} ) - # Update the state - step_errors = tool.update_state( trans, tool.inputs, step.state.inputs, step_args, - update_only=True, old_errors=old_errors ) - else: - module = step.module = module_factory.from_workflow_step( trans, step ) - state = step.state = module.decode_runtime_state( trans, step_args.pop( "tool_state" ) ) - step_errors = module.update_runtime_state( trans, state, step_args ) - if step_errors: - errors[step.id] = state.inputs["__errors__"] = step_errors - if 'run_workflow' in kwargs and not errors: - # Run each step, connecting outputs to inputs - outputs = odict() - for i, step in enumerate( workflow.steps ): - if step.type == 'tool' or step.type is None: - tool = trans.app.toolbox.tools_by_id[ step.tool_id ] - input_values = step.state.inputs - # Connect up - def callback( input, value, prefixed_name, prefixed_label ): - if isinstance( input, DataToolParameter ): - if prefixed_name in step.input_connections_by_name: - conn = step.input_connections_by_name[ prefixed_name ] - return outputs[ conn.output_step.id ][ conn.output_name ] - visit_input_values( tool.inputs, step.state.inputs, callback ) - # Execute it - outputs[ step.id ] = tool.execute( trans, step.state.inputs ) - else: - outputs[ step.id ] = step.module.execute( trans, step.state ) - - return trans.fill_template( "workflow/run_complete.mako", - workflow=stored, - outputs=outputs ) - else: - for step in workflow.steps: - if step.type == 'tool' or step.type is None: - # Restore the tool state for the step - module = module_factory.from_workflow_step( trans, step ) - # Any connected input needs to have value DummyDataset (these - # are not persisted so we need to do it every time) - module.add_dummy_datasets( connections=step.input_connections ) - # Store state with the step - step.module = module - step.state = module.state - # Error dict - if step.tool_errors: - errors[step.id] = step.tool_errors - else: - ## Non-tool specific stuff? - step.module = module_factory.from_workflow_step( trans, step ) - step.state = step.module.get_runtime_state() - # Connections by input name - step.input_connections_by_name = dict( ( conn.input_name, conn ) for conn in step.input_connections ) - # Render the form - return trans.fill_template( - "workflow/run.mako", - steps=workflow.steps, - workflow=stored, - errors=errors, - incoming=kwargs ) - - @web.expose - def configure_menu( self, trans, workflow_ids=None ): - user = trans.get_user() - if trans.request.method == "POST": - if workflow_ids is None: - workflow_ids = [] - elif type( workflow_ids ) != list: - workflow_ids = [ workflow_ids ] - sess = trans.sa_session - # This explicit remove seems like a hack, need to figure out - # how to make the association do it automatically. - for m in user.stored_workflow_menu_entries: - sess.delete( m ) - user.stored_workflow_menu_entries = [] - q = sess.query( model.StoredWorkflow ) - # To ensure id list is unique - seen_workflow_ids = set() - for id in workflow_ids: - if id in seen_workflow_ids: - continue - else: - seen_workflow_ids.add( id ) - m = model.StoredWorkflowMenuEntry() - m.stored_workflow = q.get( id ) - user.stored_workflow_menu_entries.append( m ) - sess.flush() - return trans.show_message( "Menu updated", refresh_frames=['tools'] ) - else: - user = trans.get_user() - ids_in_menu = set( [ x.stored_workflow_id for x in user.stored_workflow_menu_entries ] ) - workflows = trans.sa_session.query( model.StoredWorkflow ) \ - .filter_by( user=user, deleted=False ) \ - .order_by( desc( model.StoredWorkflow.c.update_time ) ) \ - .all() - shared_by_others = trans.sa_session \ - .query( model.StoredWorkflowUserShareAssociation ) \ - .filter_by( user=user ) \ - .filter( model.StoredWorkflow.c.deleted == False ) \ - .all() - return trans.fill_template( "workflow/configure_menu.mako", - workflows=workflows, - shared_by_others=shared_by_others, - ids_in_menu=ids_in_menu ) @web.json def json_update( self, trans ): user = trans.get_user() diff -r 6632a5d39f41 -r 802b761c5032 templates/cloud/add_credentials.mako --- a/templates/cloud/add_credentials.mako Mon Oct 26 17:04:47 2009 -0400 +++ b/templates/cloud/add_credentials.mako Tue Oct 27 15:32:44 2009 -0400 @@ -103,9 +103,9 @@ </div> %else: - In order to add credentials, desired cloud provider needs to be registered.<p/> + In order to add credentials, desired cloud provider needs to be registered first.<p/> Register <a href="${h.url_for( action='add_ec2' )}"> - <span>Amazon EC2 automatically</span></a> + <span>Amazon EC2 (us-east-1 region) automatically</span></a> or add <a href="${h.url_for( action='add_provider' )}"> <span>custom cloud provider</span></a>. diff -r 6632a5d39f41 -r 802b761c5032 templates/cloud/add_provider.mako --- a/templates/cloud/add_provider.mako Mon Oct 26 17:04:47 2009 -0400 +++ b/templates/cloud/add_provider.mako Tue Oct 27 15:32:44 2009 -0400 @@ -11,7 +11,19 @@ $("#type").change(function() { if ($(this).val() == 'ec2') { - $("#name").val("EC2"); + clear(); + $("#autofill").attr( 'disabled', true ); + $("#autofill").attr( 'checked', false ); + $("#name").val( "EC2" ); + $("#region_name").val( "us-east-1" ); + $("#region_endpoint").val( "us-east-1.ec2.amazonaws.com" ); + $("#is_secure").val("1"); + $("#debug").val(""); + $("#path").val("/"); + } + else if ($(this).val() == 'eucalyptus') { + clear(); + $("#autofill").attr( 'disabled', false ); } }); }) @@ -27,12 +39,23 @@ $("#path").val("/services/Eucalyptus"); } else { - $("#region_name").val(""); - $("#region_endpoint").val(""); - $("#is_secure").val(""); - $("#port").val(""); - $("#path").val(""); + clear(); } +} + +function clear() { + $("#name").val(""); + $("#region_name").val(""); + $("#region_endpoint").val(""); + $("#is_secure").val(""); + $("#port").val(""); + $("#proxy").val(""); + $("#proxy_port").val(""); + $("#proxy_user").val(""); + $("#proxy_pass").val(""); + $("#debug").val(""); + $("#https_connection_factory").val(""); + $("#path").val(""); } @@ -60,8 +83,9 @@ <option value="eucalyptus">Eucalyptus</option> <option value="ec2">Amazon EC2</option> </select> - <br/><input type="checkbox" id="autofill" onclick="javascript:af()"> - auto fill using Eucalyptus Public Cloud values + <br/> + <input type="checkbox" id="autofill" onclick="javascript:af()" disabled="true"> + auto fill using Eucalyptus Public Cloud values </div> %if error.has_key('type_error'): <div class="form-row-error-message">${error['type_error']}</div> @@ -90,9 +114,12 @@ %> <div class="${cls}"> <label>Region name:</label> - <div class="form-row-input"> + <div id="region_selection" class="form-row-input"> <input type="text" name="region_name" id="region_name" value="${region_name}" size="40"> </div> + %if error.has_key('name_error'): + <div class="form-row-error-message">${error['name_error']}</div> + %endif <div style="clear: both"></div> </div> @@ -107,14 +134,19 @@ <div style="clear: both"></div> </div> - <% + <% cls = "form-row" + if error.has_key('is_secure_error'): + cls += " form-row-error" %> <div class="${cls}"> <label>Is secure ('O' for False or '1' for True):</label> <div class="form-row-input"> <input type="text" name="is_secure" id="is_secure" value="${is_secure}" size="40"> </div> + %if error.has_key('is_secure_error'): + <div class="form-row-error-message">${error['is_secure_error']}; you entered: '${is_secure}'</div> + %endif <div style="clear: both"></div> </div> diff -r 6632a5d39f41 -r 802b761c5032 templates/cloud/configure_cloud.mako --- a/templates/cloud/configure_cloud.mako Mon Oct 26 17:04:47 2009 -0400 +++ b/templates/cloud/configure_cloud.mako Tue Oct 27 15:32:44 2009 -0400 @@ -35,24 +35,29 @@ } else if ( old_state=='shutting-down' && new_state=='available' ) { location.reload(true); + } + else if ( old_state=='running' && new_state=='available' ) { + location.reload(true); + } + else if ( old_state=='running' && new_state=='error' ) { + location.reload(true); + } + else if ( old_state=='pending' && new_state=='error' ) { + location.reload(true); + } + else if ( old_state=='pending' && new_state=='available' ) { + location.reload(true); } else if ( new_state=='shutting-down' || new_state=='shutting-downUCI' ) { $(elem + "-link").text( "" ); } - xmlhttp = new XMLHttpRequest(); - xmlhttp.open( "HEAD", "http://127.0.0.1:8080/admin", false ); - xmlhttp.send( null ); - //alert(xmlhttp.getAllResponseHeaders()) - console.log( "xmlhttp.readyState: %s", xmlhttp.readyState ); - console.log( "xmlhttp.status: %s", xmlhttp.status ); - if ( new_state=='running' && xmlhttp.readyState==1 ) { - console.log ("in ready statsus = 1"); - //if (xmlhttp.status==200) { - // console.log( "in status = 200" ); - // location.reload(true); - //} - } - + // In order to show 'Access Galaxy' button, the whole page needs to be refreshed. So, while Galaxy is starting, + // keep refreshing the page. Should be handled better... + else if ( $(elem+"-link").text().search('starting') && old_state=='running' && new_state=='running' ) { + //location.reload(true); + $(elem + "-link").text("Still starting..."); + } + // Update 'state' and 'time alive' fields $(elem + "-state").text( data[i].state ); if (data[i].launch_time) { @@ -184,7 +189,7 @@ context.write( '<span>Access Galaxy</span>' ) context.write( '<img src="'+h.url_for('/static/images/silk/resultset_next.png')+'" /></a></div>' ) except urllib2.URLError: - context.write( '<span>Galaxy starting...</span>' ) + context.write( 'Galaxy starting...' ) %> %endif @@ -247,8 +252,10 @@ <a onclick="document.getElementById('short').style.display = 'block'; document.getElementById('full').style.display = 'none'; return 0;" href="javascript:void(0)"> - ${str(prevInstance.error)} - </a> + error:</a><br /> + ${str(prevInstance.error)} + <p /> + <a href="${h.url_for( action='set_uci_state', id=trans.security.encode_id(prevInstance.id), state='available' )}">reset state</a> </div> %else: ${str(prevInstance.state)} @@ -278,7 +285,7 @@ You have no AWS credentials associated with your Galaxy account: <a class="action-button" href="${h.url_for( action='add' )}"> <img src="${h.url_for('/static/images/silk/add.png')}" /> - <span>Add AWS credentials</span> + <span>add credentials</span> </a> or <a href="http://aws.amazon.com/" target="_blank">

1 0

[hg] galaxy 3076: Fixed refreshing of the status of an instance ...
by Greg Von Kuster 23 Nov '09

23 Nov '09

details: http://www.bx.psu.edu/hg/galaxy/rev/022ac0f64679 changeset: 3076:022ac0f64679 user: Enis Afgan <afgane(a)gmail.com> date: Wed Oct 28 16:35:36 2009 -0400 description: Fixed refreshing of the status of an instance and assocaited Galaxy web server on main Galaxy Cloud web page. diffstat: lib/galaxy/web/controllers/cloud.py | 54 +++++++++++---------------- templates/cloud/configure_cloud.mako | 55 ++++++++++++++------------- 2 files changed, 50 insertions(+), 59 deletions(-) diffs (194 lines): diff -r 7b642483d18c -r 022ac0f64679 lib/galaxy/web/controllers/cloud.py --- a/lib/galaxy/web/controllers/cloud.py Wed Oct 28 12:28:57 2009 -0400 +++ b/lib/galaxy/web/controllers/cloud.py Wed Oct 28 16:35:36 2009 -0400 @@ -3,6 +3,7 @@ import pkg_resources pkg_resources.require( "simplejson" ) import simplejson +import urllib2 from galaxy.tools.parameters import * from galaxy.tools import DefaultToolState @@ -597,30 +598,18 @@ except ValueError: pass - # Check if Amazon EC2 has already been registered by this user - ec2_registered = trans.sa_session.query( model.CloudProvider ).filter_by( user=user, type='ec2' ).first() - if region_name or region_endpoint or name or is_secure or port or proxy or debug or path: - log.debug (" in if ") if trans.app.model.CloudProvider \ .filter_by (user=user, name=name) \ .first(): - log.debug (" in if 2 ") error['name_error'] = "A provider with that name already exist." elif name=='' or len( name ) > 255: - log.debug (" in if 3") error['name_error'] = "Provider name must be between 1 and 255 characters long." elif type=='': - log.debug (" in if 4") error['type_error'] = "Provider type must be selected." - elif ec2_registered: - log.debug (" in if 5") - error['type_error'] = "Amazon EC2 has already been registered as a provider." elif not (is_secure == 0 or is_secure == 1): - log.debug (" in if 6") error['is_secure_error'] = "Field 'is secure' can only take on a value '0' or '1'" else: - log.debug (" in else ") provider = model.CloudProvider() provider.user = user provider.type = type @@ -712,26 +701,8 @@ @web.require_login( "add Amazon EC2 provider" ) def add_ec2( self, trans ): """ Default provider setup for Amazon's EC2. """ - user = trans.get_user() - # Check if EC2 has already been registered by this user. - exists = trans.sa_session.query( model.CloudProvider ) \ - .filter_by( user=user, type='ec2' ).first() - - if not exists: - self.add_provider( trans, name='Amazon EC2', type='ec2', region_name='us-east-1', region_endpoint='us-east-1.ec2.amazonaws.com', is_secure=1, path='/' ) - return self.add( trans ) -# providers = trans.sa_session.query( model.CloudProvider ).filter_by( user=user ).all() -# return trans.fill_template( "cloud/add_credentials.mako", -# credName = '', -# providerName = '', -# accessKey = '', -# secretKey = '', -# error = {}, -# providers = providers -# ) - - trans.show_error_message( "EC2 is already registered as a cloud provider under name '%s'." % exists.name ) - return self.list( trans ) + self.add_provider( trans, name='Amazon EC2', type='ec2', region_name='us-east-1', region_endpoint='us-east-1.ec2.amazonaws.com', is_secure=1, path='/' ) + return self.add( trans ) @web.json def json_update( self, trans ): @@ -751,6 +722,25 @@ insd[uci.name] = dict return insd + @web.json + def link_update( self, trans, uci_id=0 ): + ild = {} # instance-link-dict + dict = {} + dict['uci_id'] = uci_id + try: + user = trans.get_user() + # TODO: This query can assumes only one instance under given UCI can be running (i.e., started). + inst = trans.sa_session.query( model.CloudInstance ).filter_by( user=user, uci_id=uci_id, state=uci_states.RUNNING ).first() + urllib2.urlopen( "http://" + inst.public_dns ) + dict['public_dns'] = inst.public_dns + dict['inst_id'] = inst.id + ild['data'] = dict + return ild + except urllib2.URLError: + dict['public_dns'] = False + ild['data'] = dict + return ild + ## ---- Utility methods ------------------------------------------------------- def get_provider( trans, name ): diff -r 7b642483d18c -r 022ac0f64679 templates/cloud/configure_cloud.mako --- a/templates/cloud/configure_cloud.mako Wed Oct 28 12:28:57 2009 -0400 +++ b/templates/cloud/configure_cloud.mako Wed Oct 28 16:35:36 2009 -0400 @@ -23,13 +23,13 @@ function update_state() { $.getJSON( "${h.url_for( action='json_update' )}", {}, function ( data ) { for (var i in data) { - var elem = '#' + data[i].id; + var elem = '#' + data[i].id; // Because of different list managing 'live' vs. 'available' instances, refresh entire // page on necessary state change. old_state = $(elem + "-state").text(); new_state = data[i].state; - console.log( "old_state[%d] = %s", i, old_state ); - console.log( "new_state[%d] = %s", i, new_state ); + //console.log( "old_state[%d] = %s", i, old_state ); + //console.log( "new_state[%d] = %s", i, new_state ); if ( old_state=='pending' && new_state=='running' ) { location.reload(true); } @@ -51,11 +51,27 @@ else if ( new_state=='shutting-down' || new_state=='shutting-downUCI' ) { $(elem + "-link").text( "" ); } - // In order to show 'Access Galaxy' button, the whole page needs to be refreshed. So, while Galaxy is starting, - // keep refreshing the page. Should be handled better... - else if ( $(elem+"-link").text().search('starting') && old_state=='running' && new_state=='running' ) { - //location.reload(true); - $(elem + "-link").text("Still starting..."); + + // Check if Galaxy website is accessible on given instance; if so, provide link. Otherwise, wait more. + else if ( ( $(elem+"-link").text().match('starting') || $(elem+"-link").text()=='' ) && new_state=='running' ) { + //console.log ( 'elem.text: ' + $(elem+"-link").text() ); + $.getJSON( "${h.url_for( action='link_update' )}", { uci_id: data[i].id }, function ( data ) { + for (var i in data) { + var dns = data[i].public_dns; + var uci = '#' + data[i].uci_id; + if( !dns ) { + $(uci+"-link").text( 'Galaxy starting...' ); + // http://stackoverflow.com/questions/275931/how-do-you-make-an-element-flash-… + $(uci+"-link").stop().animate({ fontSize: "14px" }, 1000).animate({ fontSize: "12px" }, 1000); + } + else { + $(uci+"-link").html( '<div align="right"><a class="action-button" href="http://'+dns+'" target="_blank">' + + '<span>Access Galaxy</span>'+ + '<img src="' + "${h.url_for( '/static/images/silk/resultset_next.png' )}" + '" /></div>' ); + $(uci+"-link").stop().animate({ fontSize: "14px" }, 1000).animate({ fontSize: "12px" }, 1000); + } + } + }); } // Update 'state' and 'time alive' fields @@ -67,6 +83,7 @@ $(elem + "-launch_time").text( "N/A" ); } } + console.log(''); }); setTimeout("update_state()", 15000); } @@ -74,6 +91,7 @@ $(function() { update_state(); }); + </script> </%def> @@ -176,25 +194,8 @@ context.write( ')' ) %> </td> - <td id="${ liveInstance.id }-link"><div align="right"> - %for j, instance in enumerate( liveInstance.instance ): - ## TODO: Once more instances will be running under the same liveInstance, additional logic will need to be added to account for that - %if instance.state == "running": - ## Wait until Galaxy server starts to show 'Access Galaxy' button - <% - import urllib2 - try: - urllib2.urlopen("http://"+instance.public_dns) - context.write( '<a class="action-button" href="http://'+instance.public_dns+'" target="_blank">' ) - context.write( '<span>Access Galaxy</span>' ) - context.write( '<img src="'+h.url_for('/static/images/silk/resultset_next.png')+'" /></a></div>' ) - except urllib2.URLError: - context.write( 'Galaxy starting...' ) - %> - - %endif - %endfor - </td> + ## Handled by JavaScript function + <td id="${ liveInstance.id }-link"></td> <td> <div popupmenu="li-${i}-popup"> <a class="action-button" confirm="Are you sure you want to stop instance '${liveInstance.name}'? Please note that this may take up to 1 minute during which time the page will not refresh." href="${h.url_for( action='stop', id=trans.security.encode_id(liveInstance.id) )}">Stop</a>

1 0

[hg] galaxy 3072: A lot of usability and UI enhancements. Curren...
by Greg Von Kuster 23 Nov '09

23 Nov '09

details: http://www.bx.psu.edu/hg/galaxy/rev/676fae1a35f9 changeset: 3072:676fae1a35f9 user: Enis Afgan <afgane(a)gmail.com> date: Fri Oct 23 17:41:35 2009 -0400 description: A lot of usability and UI enhancements. Current state of automatic state upadtes causes some inconsistencies though. diffstat: lib/galaxy/cloud/__init__.py | 27 +- lib/galaxy/cloud/providers/ec2.py | 22 +- lib/galaxy/cloud/providers/eucalyptus.py | 67 ++- lib/galaxy/model/__init__.py | 6 + lib/galaxy/model/mapping.py | 41 ++- lib/galaxy/model/migrate/versions/0014_cloud_tables.py | 59 +++- lib/galaxy/web/controllers/cloud.py | 351 ++++++++++++--------- templates/cloud/add_credentials.mako | 27 +- templates/cloud/add_provider.mako | 225 ++++++++++++++ templates/cloud/configure_cloud.mako | 98 ++++- templates/cloud/view.mako | 81 ++++- templates/cloud/view_usage.mako | 47 ++ 12 files changed, 810 insertions(+), 241 deletions(-) diffs (1661 lines): diff -r 7f3054622e87 -r 676fae1a35f9 lib/galaxy/cloud/__init__.py --- a/lib/galaxy/cloud/__init__.py Wed Oct 21 11:40:51 2009 -0400 +++ b/lib/galaxy/cloud/__init__.py Fri Oct 23 17:41:35 2009 -0400 @@ -519,16 +519,25 @@ vol = model.CloudStore.filter( model.CloudStore.c.volume_id == vol_id ).first() vol.i_id = instance_id vol.flush() + + def set_error( self, error ): + """ + Sets error field of given UCI in local Galaxy database + """ + uci = model.UCI.get( self.uci_id ) + uci.refresh() + uci.error = error + uci.flush() # --------- Getter methods ----------------- - def get_provider_name( self ): - """ Returns name of cloud provider associated with given UCI. """ + def get_provider_type( self ): + """ Returns type of cloud provider associated with given UCI. """ uci = model.UCI.get( self.uci_id ) uci.refresh() - cred_id = uci.credentials_id - cred = model.CloudUserCredentials.get( cred_id ) - return cred.provider_name +# cred_id = uci.credentials_id +# cred = model.CloudUserCredentials.get( cred_id ) + return uci.credentials.provider.type def get_instances_indexes( self, state=None ): """ @@ -652,6 +661,12 @@ uci.refresh() return uci + def get_provider( self ): + """ Returns database object of cloud provider associated with credentials of given UCI. """ + uci = model.UCI.get( self.uci_id ) + uci.refresh() + return uci.credentials.provider + def uci_launch_time_set( self ): uci = model.UCI.get( self.uci_id ) uci.refresh() @@ -1021,7 +1036,7 @@ def put( self, uci_wrapper ): """ Put given request for UCI manipulation into provider's request queue.""" # log.debug( "Adding UCI '%s' manipulation request into cloud manager's queue." % uci_wrapper.name ) - self.cloud_provider[uci_wrapper.get_provider_name()].put( uci_wrapper ) + self.cloud_provider[uci_wrapper.get_provider_type()].put( uci_wrapper ) diff -r 7f3054622e87 -r 676fae1a35f9 lib/galaxy/cloud/providers/ec2.py --- a/lib/galaxy/cloud/providers/ec2.py Wed Oct 21 11:40:51 2009 -0400 +++ b/lib/galaxy/cloud/providers/ec2.py Fri Oct 23 17:41:35 2009 -0400 @@ -51,7 +51,7 @@ """ STOP_SIGNAL = object() def __init__( self, app ): - self.name = "ec2" + self.type = "ec2" # cloud provider type (e.g., ec2, eucalyptus, opennebula) self.zone = "us-east-1a" self.key_pair = "galaxy-keypair" self.queue = Queue() @@ -73,7 +73,7 @@ uci_wrapper = self.queue.get() # uci = uci_wrapper.get_uci() - log.debug( '[%d] uci name: %s' % ( cnt, uci_wrapper.get_name() ) ) + log.debug( '[%d] uci type: %s' % ( cnt, uci_wrapper.get_name() ) ) uci_state = uci_wrapper.get_state() if uci_state is self.STOP_SIGNAL: return @@ -118,6 +118,7 @@ uci_wrapper.set_key_pair( inst, kp.name, kp.material ) else: log.error( "EC2 response error: '%s'" % e ) + uci_wrapper.set_error( "EC2 response error while creating key pair: " + e ) return kp.name @@ -194,6 +195,7 @@ else: uci_wrapper.change_state( uci_state=uci_states.ERROR ) uci_wrapper.set_store_status( vol.id, uci_states.ERROR ) + uci_wrapper.set_error( "Volume '%s' not found by cloud provider after being created" % vol.id ) def deleteUCI( self, uci_wrapper ): """ @@ -226,6 +228,8 @@ log.error( "Deleting following volume(s) failed: %s. However, these volumes were successfully deleted: %s. \ MANUAL intervention and processing needed." % ( failedList, deletedList ) ) uci_wrapper.change_state( uci_state=uci_state.ERROR ) + uci_wrapper.set_error( "Deleting following volume(s) failed: "+failedList+". However, these volumes were successfully deleted: "+deletedList+". \ + MANUAL intervention and processing needed." ) def addStorageToUCI( self, name ): """ Adds more storage to specified UCI @@ -291,7 +295,7 @@ uci_wrapper.set_instance_id( i_index, i_id ) s = reservation.instances[0].state uci_wrapper.change_state( s, i_id, s ) - log.debug( "Instance of UCI '%s' started, current state: %s" % ( uci_wrapper.get_name(), uci_wrapper.get_state() ) ) + log.debug( "Instance of UCI '%s' started, current state: '%s'" % ( uci_wrapper.get_name(), uci_wrapper.get_state() ) ) @@ -429,20 +433,20 @@ Reason behind this method is to sync state of local DB and real-world resources """ log.debug( "Running general status update for EC2 UCIs..." ) - instances = model.CloudInstance.filter( or_( model.CloudInstance.c.state==instance_states.RUNNING, #"running", - model.CloudInstance.c.state==instance_states.PENDING, #"pending", + instances = model.CloudInstance.filter( or_( model.CloudInstance.c.state==instance_states.RUNNING, + model.CloudInstance.c.state==instance_states.PENDING, model.CloudInstance.c.state==instance_states.SHUTTING_DOWN ) ).all() for inst in instances: - if self.name == inst.uci.credentials.provider_name: - log.debug( "[%s] Running general status update on instance '%s'" % ( inst.uci.credentials.provider_name, inst.instance_id ) ) + if self.type == inst.uci.credentials.provider.type: + log.debug( "[%s] Running general status update on instance '%s'" % ( inst.uci.credentials.provider.type, inst.instance_id ) ) self.updateInstance( inst ) stores = model.CloudStore.filter( or_( model.CloudStore.c.status==store_states.IN_USE, model.CloudStore.c.status==store_states.CREATING, model.CloudStore.c.status==None ) ).all() for store in stores: - if self.name == store.uci.credentials.provider_name: - log.debug( "[%s] Running general status update on store '%s'" % ( store.uci.credentials.provider_name, store.volume_id ) ) + if self.type == store.uci.credentials.provider.type: + log.debug( "[%s] Running general status update on store '%s'" % ( store.uci.credentials.provider.type, store.volume_id ) ) self.updateStore( store ) def updateInstance( self, inst ): diff -r 7f3054622e87 -r 676fae1a35f9 lib/galaxy/cloud/providers/eucalyptus.py --- a/lib/galaxy/cloud/providers/eucalyptus.py Wed Oct 21 11:40:51 2009 -0400 +++ b/lib/galaxy/cloud/providers/eucalyptus.py Fri Oct 23 17:41:35 2009 -0400 @@ -51,7 +51,7 @@ """ STOP_SIGNAL = object() def __init__( self, app ): - self.name = "eucalyptus" + self.type = "eucalyptus" # cloud provider type (e.g., ec2, eucalyptus, opennebula) self.zone = "epc" self.key_pair = "galaxy-keypair" self.queue = Queue() @@ -73,7 +73,7 @@ uci_wrapper = self.queue.get() # uci = uci_wrapper.get_uci() - log.debug( '[%d] uci name: %s' % ( cnt, uci_wrapper.get_name() ) ) + log.debug( '[%d] uci type: %s' % ( cnt, uci_wrapper.get_name() ) ) uci_state = uci_wrapper.get_state() if uci_state is self.STOP_SIGNAL: return @@ -95,17 +95,21 @@ Establishes eucalyptus cloud connection using user's credentials associated with given UCI """ log.debug( '##### Establishing eucalyptus cloud connection' ) - # Eucalyptus Public Cloud - # TODO: Add option in Galaxy config file to specify these values (i.e., for locally managed Eucalyptus deployments) - euca_region = RegionInfo( None, "eucalyptus", "mayhem9.cs.ucsb.edu" ) - conn = EC2Connection( aws_access_key_id=uci_wrapper.get_access_key(), aws_secret_access_key=uci_wrapper.get_secret_key(), is_secure=False, port=8773, region=euca_region, path="/services/Eucalyptus" ) + provider = uci_wrapper.get_provider() + euca_region = RegionInfo( None, provider.region_name, provider.region_endpoint ) + conn = EC2Connection( aws_access_key_id=uci_wrapper.get_access_key(), + aws_secret_access_key=uci_wrapper.get_secret_key(), + is_secure=provider.is_secure, + port=provider.port, + region=euca_region, + path=provider.path ) return conn def set_keypair( self, uci_wrapper, conn ): """ Generate keypair using user's default credentials """ - log.debug( "Getting user's keypair" ) + log.debug( "Getting user's keypair: '%s'" % self.key_pair ) kp = conn.get_key_pair( self.key_pair ) instances = uci_wrapper.get_instances_indexes() @@ -128,7 +132,8 @@ TODO: Dummy method - need to implement logic For valid sizes, see http://aws.amazon.com/ec2/instance-types/ """ - return model.CloudImage.filter( model.CloudImage.table.c.id==1 ).first().image_id + log.debug( "image id: '%s'" % model.CloudImage.get( 1 ).image_id ) + return model.CloudImage.get( 1 ).image_id # def get_instances( self, uci ): # """ @@ -211,6 +216,8 @@ log.error( "Deleting following volume(s) failed: %s. However, these volumes were successfully deleted: %s. \ MANUAL intervention and processing needed." % ( failedList, deletedList ) ) uci_wrapper.change_state( uci_state=uci_states.ERROR ) + uci_wrapper.set_error( "Deleting following volume(s) failed: "+failedList+". However, these volumes were successfully deleted: "+deletedList+". \ + MANUAL intervention and processing needed." ) def addStorageToUCI( self, name ): """ Adds more storage to specified UCI """ @@ -251,7 +258,7 @@ # for i in range( len( sgs ) ): # if sgs[i].name == "galaxy": # sg.append( sgs[i] ) - # break # only 1 security group w/ this name can exist, so continue + # break # only 1 security group w/ this type can exist, so continue log.debug( "***** Starting UCI instance '%s'" % uci_wrapper.get_name() ) log.debug( 'Using following command: conn.run_instances( image_id=%s, key_name=%s )' % ( mi_id, uci_wrapper.get_key_pair_name( i_index ) ) ) @@ -267,7 +274,7 @@ uci_wrapper.set_instance_id( i_index, i_id ) s = reservation.instances[0].state uci_wrapper.change_state( s, i_id, s ) - log.debug( "Instance of UCI '%s' started, current state: %s" % ( uci_wrapper.get_name(), uci_wrapper.get_state() ) ) + log.debug( "Instance of UCI '%s' started, current state: '%s'" % ( uci_wrapper.get_name(), uci_wrapper.get_state() ) ) @@ -310,14 +317,9 @@ # Get all instances associated with given UCI il = uci_wrapper.get_instances_ids() # instance list -# log.debug( 'List of instances being terminated: %s' % il ) + log.debug( 'List of instances being terminated: %s' % il ) rl = conn.get_all_instances( il ) # Reservation list associated with given instances - -# tState = conn.terminate_instances( il ) -# # TODO: Need to update instance stop time (for all individual instances) -# stop_time = datetime.utcnow() -# uci_wrapper.set_stop_time( stop_time ) - + # Initiate shutdown of all instances under given UCI cnt = 0 stopped = [] @@ -406,20 +408,20 @@ Reason behind this method is to sync state of local DB and real-world resources """ log.debug( "Running general status update for EPC UCIs..." ) - instances = model.CloudInstance.filter( or_( model.CloudInstance.c.state==instance_states.RUNNING, #"running", - model.CloudInstance.c.state==instance_states.PENDING, #"pending", + instances = model.CloudInstance.filter( or_( model.CloudInstance.c.state==instance_states.RUNNING, + model.CloudInstance.c.state==instance_states.PENDING, model.CloudInstance.c.state==instance_states.SHUTTING_DOWN ) ).all() for inst in instances: - if self.name == inst.uci.credentials.provider_name: - log.debug( "[%s] Running general status update on instance '%s'" % ( inst.uci.credentials.provider_name, inst.instance_id ) ) + if self.type == inst.uci.credentials.provider.type: + log.debug( "[%s] Running general status update on instance '%s'" % ( inst.uci.credentials.provider.type, inst.instance_id ) ) self.updateInstance( inst ) stores = model.CloudStore.filter( or_( model.CloudStore.c.status==store_states.IN_USE, model.CloudStore.c.status==store_states.CREATING, model.CloudStore.c.status==None ) ).all() for store in stores: - if self.name == store.uci.credentials.provider_name: - log.debug( "[%s] Running general status update on store '%s'" % ( store.uci.credentials.provider_name, store.volume_id ) ) + if self.type == store.uci.credentials.provider.type: + log.debug( "[%s] Running general status update on store '%s'" % ( store.uci.credentials.provider.type, store.volume_id ) ) self.updateStore( store ) def updateInstance( self, inst ): @@ -431,8 +433,13 @@ a_key = uci.credentials.access_key s_key = uci.credentials.secret_key # Get connection - euca_region = RegionInfo( None, "eucalyptus", "mayhem9.cs.ucsb.edu" ) - conn = EC2Connection( aws_access_key_id=a_key, aws_secret_access_key=s_key, is_secure=False, port=8773, region=euca_region, path="/services/Eucalyptus" ) + euca_region = RegionInfo( None, uci.credentials.provider.region_name, uci.credentials.provider.region_endpoint ) + conn = EC2Connection( aws_access_key_id=a_key, + aws_secret_access_key=s_key, + is_secure=uci.credentials.provider.is_secure, + port=uci.credentials.provider.port, + region=euca_region, + path=uci.credentials.provider.path ) # Get reservations handle for given instance rl= conn.get_all_instances( [inst.instance_id] ) # Because EPC deletes references to reservations after a short while after instances have terminated, getting an empty list as a response to a query @@ -475,9 +482,13 @@ a_key = uci.credentials.access_key s_key = uci.credentials.secret_key # Get connection - euca_region = RegionInfo( None, "eucalyptus", "mayhem9.cs.ucsb.edu" ) - conn = EC2Connection( aws_access_key_id=a_key, aws_secret_access_key=s_key, is_secure=False, port=8773, region=euca_region, path="/services/Eucalyptus" ) - # Get reservations handle for given store + euca_region = RegionInfo( None, uci.credentials.provider.region_name, uci.credentials.provider.region_endpoint ) + conn = EC2Connection( aws_access_key_id=a_key, + aws_secret_access_key=s_key, + is_secure=uci.credentials.provider.is_secure, + port=uci.credentials.provider.port, + region=euca_region, + path=uci.credentials.provider.path )# Get reservations handle for given store vl = conn.get_all_volumes( [store.volume_id] ) # log.debug( "Store '%s' vl: '%s'" % ( store.volume_id, vl ) ) # Update store status in local DB with info from cloud provider diff -r 7f3054622e87 -r 676fae1a35f9 lib/galaxy/model/__init__.py --- a/lib/galaxy/model/__init__.py Wed Oct 21 11:40:51 2009 -0400 +++ b/lib/galaxy/model/__init__.py Fri Oct 23 17:41:35 2009 -0400 @@ -963,6 +963,12 @@ self.user = None self.size = None self.availability_zone = None + +class CloudProvider( object ): + def __init__( self ): + self.id = None + self.user = None + self.type = None class CloudUserCredentials( object ): def __init__( self ): diff -r 7f3054622e87 -r 676fae1a35f9 lib/galaxy/model/mapping.py --- a/lib/galaxy/model/mapping.py Wed Oct 21 11:40:51 2009 -0400 +++ b/lib/galaxy/model/mapping.py Fri Oct 23 17:41:35 2009 -0400 @@ -395,14 +395,15 @@ Column( "state", TEXT ) ) """ UserConfiguredInstance (UCI) table """ -UCI.table = Table( "uci", metadata, +UCI.table = Table( "cloud_uci", metadata, Column( "id", Integer, primary_key=True ), Column( "create_time", DateTime, default=now ), Column( "update_time", DateTime, default=now, onupdate=now ), Column( "user_id", Integer, ForeignKey( "galaxy_user.id" ), index=True, nullable=False ), - Column( "credentials_id", Integer, ForeignKey( "cloud_user_credentials.id" ), index=True, nullable=False ), + Column( "credentials_id", Integer, ForeignKey( "cloud_user_credentials.id" ), index=True ), Column( "name", TEXT ), Column( "state", TEXT ), + Column( "error", TEXT ), Column( "total_size", Integer ), Column( "launch_time", DateTime ) ) @@ -413,12 +414,13 @@ Column( "launch_time", DateTime ), Column( "stop_time", DateTime ), Column( "user_id", Integer, ForeignKey( "galaxy_user.id" ), index=True, nullable=False ), - Column( "uci_id", Integer, ForeignKey( "uci.id" ), index=True ), + Column( "uci_id", Integer, ForeignKey( "cloud_uci.id" ), index=True ), Column( "type", TEXT ), Column( "reservation_id", TEXT ), Column( "instance_id", TEXT ), Column( "mi_id", TEXT, ForeignKey( "cloud_image.image_id" ), index=True, nullable=False ), Column( "state", TEXT ), + Column( "error", TEXT ), Column( "public_dns", TEXT ), Column( "private_dns", TEXT ), Column( "keypair_name", TEXT ), @@ -431,7 +433,7 @@ Column( "update_time", DateTime, default=now, onupdate=now ), Column( "attach_time", DateTime ), Column( "user_id", Integer, ForeignKey( "galaxy_user.id" ), index=True, nullable=False ), - Column( "uci_id", Integer, ForeignKey( "uci.id" ), index=True, nullable=False ), + Column( "uci_id", Integer, ForeignKey( "cloud_uci.id" ), index=True, nullable=False ), Column( "volume_id", TEXT ), Column( "size", Integer, nullable=False ), Column( "availability_zone", TEXT ), @@ -440,6 +442,27 @@ Column( "device", TEXT ), Column( "space_consumed", Integer ) ) +CloudProvider.table = Table( "cloud_provider", metadata, + Column( "id", Integer, primary_key=True ), + Column( "create_time", DateTime, default=now ), + Column( "update_time", DateTime, default=now, onupdate=now ), + Column( "user_id", Integer, ForeignKey( "galaxy_user.id" ), index=True, nullable=False ), + Column( "type", TEXT, nullable=False ), + Column( "name", TEXT ), + Column( "region_connection", TEXT ), + Column( "region_name", TEXT ), + Column( "region_endpoint", TEXT ), + Column( "is_secure", Boolean ), + Column( "host", TEXT ), + Column( "port", Integer ), + Column( "proxy", TEXT ), + Column( "proxy_port", TEXT ), + Column( "proxy_user", TEXT ), + Column( "proxy_pass", TEXT ), + Column( "debug", Integer ), + Column( "https_connection_factory", TEXT ), + Column( "path", TEXT ) ) + CloudUserCredentials.table = Table( "cloud_user_credentials", metadata, Column( "id", Integer, primary_key=True ), Column( "create_time", DateTime, default=now ), @@ -449,7 +472,8 @@ Column( "access_key", TEXT ), Column( "secret_key", TEXT ), Column( "defaultCred", Boolean, default=False ), - Column( "provider_name", TEXT ) ) + Column( "provider_id", Integer, ForeignKey( "cloud_provider.id" ), index=True, nullable=False ) ) + # *************************************************************************** StoredWorkflow.table = Table( "stored_workflow", metadata, @@ -978,8 +1002,13 @@ i=relation( CloudInstance ) ) ) +assign_mapper( context, CloudProvider, CloudProvider.table, + properties=dict( user=relation( User ) + ) ) + assign_mapper( context, CloudUserCredentials, CloudUserCredentials.table, - properties=dict( user=relation( User) + properties=dict( user=relation( User), + provider=relation( CloudProvider ) ) ) # ^^^^^^^^^^^^^^^ End cloud table mappings ^^^^^^^^^^^^^^^^^^ diff -r 7f3054622e87 -r 676fae1a35f9 lib/galaxy/model/migrate/versions/0014_cloud_tables.py --- a/lib/galaxy/model/migrate/versions/0014_cloud_tables.py Wed Oct 21 11:40:51 2009 -0400 +++ b/lib/galaxy/model/migrate/versions/0014_cloud_tables.py Fri Oct 23 17:41:35 2009 -0400 @@ -20,14 +20,15 @@ Column( "manifest", TEXT ), Column( "state", TEXT ) ) -UCI_table = Table( "uci", metadata, +UCI_table = Table( "cloud_uci", metadata, Column( "id", Integer, primary_key=True ), Column( "create_time", DateTime, default=now ), Column( "update_time", DateTime, default=now, onupdate=now ), Column( "user_id", Integer, ForeignKey( "galaxy_user.id" ), index=True, nullable=False ), - Column( "credentials_id", Integer, ForeignKey( "cloud_user_credentials.id" ), index=True, nullable=False ), + Column( "credentials_id", Integer, ForeignKey( "cloud_user_credentials.id" ), index=True ), Column( "name", TEXT ), Column( "state", TEXT ), + Column( "error", TEXT ), Column( "total_size", Integer ), Column( "launch_time", DateTime ) ) @@ -38,12 +39,13 @@ Column( "launch_time", DateTime ), Column( "stop_time", DateTime ), Column( "user_id", Integer, ForeignKey( "galaxy_user.id" ), index=True, nullable=False ), - Column( "uci_id", Integer, ForeignKey( "uci.id" ), index=True, nullable=False ), + Column( "uci_id", Integer, ForeignKey( "uci.id" ), index=True ), Column( "type", TEXT ), Column( "reservation_id", TEXT ), Column( "instance_id", TEXT ), - Column( "mi_id", TEXT, ForeignKey( "cloud_image.image_id" ), index=True ), + Column( "mi_id", TEXT, ForeignKey( "cloud_image.image_id" ), index=True, nullable=False ), Column( "state", TEXT ), + Column( "error", TEXT ), Column( "public_dns", TEXT ), Column( "private_dns", TEXT ), Column( "keypair_name", TEXT ), @@ -74,7 +76,28 @@ Column( "access_key", TEXT ), Column( "secret_key", TEXT ), Column( "defaultCred", Boolean, default=False ), - Column( "provider_name", TEXT ) ) + Column( "provider_id", Integer, ForeignKey( "cloud_provider.id" ), index=True, nullable=False ) ) + +CloudProvider_table = Table( "cloud_provider", metadata, + Column( "id", Integer, primary_key=True ), + Column( "create_time", DateTime, default=now ), + Column( "update_time", DateTime, default=now, onupdate=now ), + Column( "user_id", Integer, ForeignKey( "galaxy_user.id" ), index=True, nullable=False ), + Column( "type", TEXT, nullable=False ), + Column( "name", TEXT ), + Column( "region_connection", TEXT ), + Column( "region_name", TEXT ), + Column( "region_endpoint", TEXT ), + Column( "is_secure", Boolean ), + Column( "host", TEXT ), + Column( "port", Integer ), + Column( "proxy", TEXT ), + Column( "proxy_port", TEXT ), + Column( "proxy_user", TEXT ), + Column( "proxy_pass", TEXT ), + Column( "debug", Integer ), + Column( "https_connection_factory", TEXT ), + Column( "path", TEXT ) ) def upgrade(): metadata.reflect() @@ -86,12 +109,22 @@ UCI_table.create() except Exception, e: log.debug( "Creating UCI table failed. Table probably exists already." ) - CloudInstance_table.create() - CloudStore_table.create() + try: + CloudInstance_table.create() + except Exception, e: + log.debug( "Creating cloud_instance table failed. Table probably exists already." ) + try: + CloudStore_table.create() + except Exception: + log.debug( "Creating cloud_store table failed. Table probably exists already." ) try: CloudUserCredentials_table.create() except Exception, e: log.debug( "Creating cloud_image table failed. Table probably exists already." ) + try: + CloudProvider_table.create() + except Exception, e: + log.debug( "Creating cloud_provider table failed. Table probably exists already." ) def downgrade(): metadata.reflect() @@ -102,12 +135,14 @@ log.debug( "Dropping cloud_image table failed: %s" % str( e ) ) try: - CloudInstance_table.drop() + log.debug( "Would drop cloud_instance table." ) +# CloudInstance_table.drop() except Exception, e: log.debug( "Dropping cloud_instance table failed: %s" % str( e ) ) try: - CloudStore_table.drop() + log.debug( "Would drop cloud_store table." ) +# CloudStore_table.drop() except Exception, e: log.debug( "Dropping cloud_store table failed: %s" % str( e ) ) @@ -122,6 +157,12 @@ # UCI_table.drop() except Exception, e: log.debug( "Dropping UCI table failed: %s" % str( e ) ) + + try: +# log.debug( "Would drop cloud_provider table." ) + CloudProvider_table.drop() + except Exception, e: + log.debug( "Dropping cloud_provider table failed: %s" % str( e ) ) diff -r 7f3054622e87 -r 676fae1a35f9 lib/galaxy/web/controllers/cloud.py --- a/lib/galaxy/web/controllers/cloud.py Wed Oct 21 11:40:51 2009 -0400 +++ b/lib/galaxy/web/controllers/cloud.py Fri Oct 23 17:41:35 2009 -0400 @@ -77,7 +77,7 @@ cloudCredentials = trans.sa_session.query( model.CloudUserCredentials ) \ .filter_by( user=user ) \ - .order_by( desc( model.CloudUserCredentials.c.update_time ) ) \ + .order_by( desc( model.CloudUserCredentials.c.name ) ) \ .all() liveInstances = trans.sa_session.query( model.UCI ) \ @@ -115,6 +115,13 @@ "start up and then refresh this page. A button to connect to the instance will then appear alongside " "instance description." ) +# log.debug( "provider.is_secure: '%s'" % trans.sa_session.query( model.CloudProvider).filter_by(id=1).first().is_secure ) +# trans.sa_session.query( model.CloudProvider).filter_by(id=1).first().is_secure=False +# trans.sa_session.flush() +# log.debug( "provider.is_secure: '%s'" % trans.sa_session.query( model.CloudProvider).filter_by(id=1).first().is_secure ) + +# log.debug( "image: '%s'" % model.CloudImage.is_secure ) + return trans.fill_template( "cloud/configure_cloud.mako", cloudCredentials = cloudCredentials, liveInstances = liveInstances, @@ -146,13 +153,12 @@ """ Start a new cloud resource instance """ - # TODO: Add choice of instance type before starting one - #if type: user = trans.get_user() - mi = get_mi( trans, type ) uci = get_uci( trans, id ) + mi = get_mi( trans, uci, type ) stores = get_stores( trans, uci ) -# log.debug(self.app.config.job_working_directory) + # Ensure instance is not already running (or related state) and store relevant data + # into DB to initiate instance startup by cloud manager if ( len(stores) is not 0 ) and \ ( uci.state != uci_states.SUBMITTED ) and \ ( uci.state != uci_states.SUBMITTED_UCI ) and \ @@ -166,116 +172,45 @@ instance.uci = uci instance.availability_zone = stores[0].availability_zone # Bc. all EBS volumes need to be in the same avail. zone, just check 1st instance.type = type -# instance.keypair_name = get_keypair_name( trans ) -# conn = get_connection( trans ) -# log.debug( '***** Setting up security group' ) - # If not existent, setup galaxy security group -# try: -# gSecurityGroup = conn.create_security_group('galaxy', 'Security group for Galaxy.') -# gSecurityGroup.authorize( 'tcp', 80, 80, '0.0.0.0/0' ) # Open HTTP port -# gSecurityGroup.authorize( 'tcp', 22, 22, '0.0.0.0/0' ) # Open SSH port -# except: -# pass -# sgs = conn.get_all_security_groups() -# for i in range( len( sgs ) ): -# if sgs[i].name == "galaxy": -# sg.append( sgs[i] ) -# break # only 1 security group w/ this name can exist, so continue - -# log.debug( '***** Starting an instance' ) -# log.debug( 'Using following command: conn.run_instances( image_id=%s, key_name=%s )' % ( instance.image.image_id, instance.keypair_name ) ) -# reservation = conn.run_instances( image_id=instance.image.image_id, key_name=instance.keypair_name ) - #reservation = conn.run_instances( image_id=instance.image, key_name=instance.keypair_name, security_groups=['galaxy'], instance_type=instance.type, placement=instance.availability_zone ) -# instance.launch_time = datetime.utcnow() -# uci.launch_time = instance.launch_time -# instance.reservation_id = str( reservation ).split(":")[1] -# instance.instance_id = str( reservation.instances[0]).split(":")[1] -# instance.state = "pending" -# instance.state = reservation.instances[0].state uci.state = uci_states.SUBMITTED_UCI - # Persist session = trans.sa_session session.save_or_update( instance ) session.save_or_update( uci ) session.flush() - + # Log trans.log_event ("User initiated starting of cloud instance '%s'." % uci.name ) trans.set_message( "Galaxy instance started. NOTE: Please wait about 3-5 minutes for the instance to " "start up and then refresh this page. A button to connect to the instance will then appear alongside " "instance description." ) - time.sleep(1) # Wait for initial update to occur to avoid immediate page reload return self.list( trans ) trans.show_error_message( "Cannot start instance that is in state '%s'." % uci.state ) return self.list( trans ) - -# return trans.show_form( -# web.FormBuilder( web.url_for(), "Start instance size", submit_text="Start" ) -# .add_input( "radio","Small","size", value='small' ) -# .add_input( "radio","Medium","size", value='medium' ) ) - @web.expose @web.require_login( "stop Galaxy cloud instance" ) def stop( self, trans, id ): """ - Stop a cloud UCI instance. This implies stopping Galaxy server and disconnecting/unmounting relevant file system(s). + Stop a cloud UCI instance. """ uci = get_uci( trans, id ) - uci.state = uci_states.SHUTTING_DOWN_UCI - session = trans.sa_session -# session.save_or_update( stores ) - session.save_or_update( uci ) - session.flush() - trans.log_event( "User stopped cloud instance '%s'" % uci.name ) - trans.set_message( "Galaxy instance '%s' stopped." % uci.name ) + if ( uci.state != uci_states.DELETING ) and \ + ( uci.state != uci_states.DELETING_UCI ) and \ + ( uci.state != uci_states.ERROR ) and \ + ( uci.state != uci_states.SHUTTING_DOWN_UCI ) and \ + ( uci.state != uci_states.SHUTTING_DOWN ) and \ + ( uci.state != uci_states.AVAILABLE ): + uci.state = uci_states.SHUTTING_DOWN_UCI + session = trans.sa_session + session.save_or_update( uci ) + session.flush() + trans.log_event( "User stopped cloud instance '%s' (id: %s)" % ( uci.name, uci.id ) ) + trans.set_message( "Stopping of Galaxy instance '%s' initiated." % uci.name ) + + return self.list( trans ) -# dbInstances = get_instances( trans, uci ) #TODO: handle list! -# -# conn = get_connection( trans ) -# # Get actual cloud instance object -# cloudInstance = get_cloud_instance( conn, dbInstances.instance_id ) -# -# # TODO: Detach persistent storage volume(s) from instance and update volume data in local database -# stores = get_stores( trans, uci ) -# for i, store in enumerate( stores ): -# log.debug( "Detaching volume '%s' to instance '%s'." % ( store.volume_id, dbInstances.instance_id ) ) -# mntDevice = store.device -# volStat = None -## Detaching volume does not work with Eucalyptus Public Cloud, so comment it out -## try: -## volStat = conn.detach_volume( store.volume_id, dbInstances.instance_id, mntDevice ) -## except: -## log.debug ( 'Error detaching volume; still going to try and stop instance %s.' % dbInstances.instance_id ) -# store.attach_time = None -# store.device = None -# store.i_id = None -# store.status = volStat -# log.debug ( '***** volume status: %s' % volStat ) -# -# -# # Stop the instance and update status in local database -# cloudInstance.stop() -# dbInstances.stop_time = datetime.utcnow() -# while cloudInstance.state != 'terminated': -# log.debug( "Stopping instance %s state; current state: %s" % ( str( cloudInstance ).split(":")[1], cloudInstance.state ) ) -# time.sleep(3) -# cloudInstance.update() -# dbInstances.state = cloudInstance.state -# -# # Reset relevant UCI fields -# uci.state = 'available' -# uci.launch_time = None -# -# # Persist -# session = trans.sa_session -## session.save_or_update( stores ) -# session.save_or_update( dbInstances ) # TODO: Is this going to work w/ multiple instances stored in dbInstances variable? -# session.save_or_update( uci ) -# session.flush() -# trans.log_event( "User stopped cloud instance '%s'" % uci.name ) -# trans.set_message( "Galaxy instance '%s' stopped." % uci.name ) + trans.show_error_message( "Cannot stop instance that is in state '%s'." % uci.state ) return self.list( trans ) @web.expose @@ -290,30 +225,10 @@ if ( uci.state != uci_states.DELETING_UCI ) and ( uci.state != uci_states.DELETING ) and ( uci.state != uci_states.ERROR ): name = uci.name uci.state = uci_states.DELETING_UCI - # dbInstances = get_instances( trans, uci ) #TODO: handle list! - # - # conn = get_connection( trans ) session = trans.sa_session - # - # # Delete volume(s) associated with given uci - # stores = get_stores( trans, uci ) - # for i, store in enumerate( stores ): - # log.debug( "Deleting volume '%s' that is associated with UCI '%s'." % ( store.volume_id, uci.name ) ) - # volStat = None - # try: - # volStat = conn.delete_volume( store.volume_id ) - # except: - # log.debug ( 'Error deleting volume %s' % store.volume_id ) - # - # if volStat: - # session.delete( store ) - # - # # Delete UCI from table - # uciName = uci.name # Store name for logging - # session.delete( uci ) - + session.save_or_update( uci ) session.flush() - trans.log_event( "User deleted cloud instance '%s'" % name ) + trans.log_event( "User marked cloud instance '%s' for deletion." % name ) trans.set_message( "Galaxy instance '%s' marked for deletion." % name ) return self.list( trans ) @@ -332,7 +247,19 @@ @web.expose @web.require_login( "use Galaxy cloud" ) - def configureNew( self, trans, instanceName='', credName='', volSize='', zone=''): + def usageReport( self, trans ): + user = trans.get_user() + + prevInstances = trans.sa_session.query( model.CloudInstance ) \ + .filter_by( user=user, state=instance_states.TERMINATED ) \ + .order_by( desc( model.CloudInstance.c.update_time ) ) \ + .all() + + return trans.fill_template( "cloud/view_usage.mako", prevInstances = prevInstances ) + + @web.expose + @web.require_login( "use Galaxy cloud" ) + def configureNew( self, trans, instanceName='', credName='', volSize='', zone='' ): """ Configure and add new cloud instance to user's instance pool """ @@ -346,16 +273,19 @@ providersToZones = {} for storedCred in storedCreds: - if storedCred.provider_name == 'ec2': + if storedCred.provider.type == 'ec2': ec2_zones = ['us-east-1a', 'us-east-1b', 'us-east-1c', 'us-east-1d'] providersToZones[storedCred.name] = ec2_zones - elif storedCred.provider_name == 'eucalyptus': + elif storedCred.provider.type == 'eucalyptus': providersToZones[storedCred.name] = ['epc'] if instanceName: # Create new user configured instance try: - if trans.app.model.UCI.filter( and_( trans.app.model.UCI.table.c.name==instanceName, trans.app.model.UCI.table.c.state!=uci_states.DELETED ) ).first(): + if trans.app.model.UCI \ + .filter_by (user=user) \ + .filter( and_( trans.app.model.UCI.table.c.name==instanceName, trans.app.model.UCI.table.c.state!=uci_states.DELETED ) ) \ + .first(): error['inst_error'] = "An instance with that name already exist." elif instanceName=='' or len( instanceName ) > 255: error['inst_error'] = "Instance name must be between 1 and 255 characters long." @@ -393,7 +323,6 @@ # Log and display the management page trans.log_event( "User configured new cloud instance" ) trans.set_message( "New Galaxy instance '%s' configured. Once instance status shows 'available' you will be able to start the instance." % instanceName ) - time.sleep(1) # Wait for initial update to occur to avoid immediate page reload return self.list( trans ) except ValueError: vol_error = "Volume size must be specified as an integer value only, between 1 and 1000." @@ -460,8 +389,9 @@ trans.set_message( "Credentials renamed to '%s'." % new_name ) return self.list( trans ) else: - return form( url_for( id=trans.security.encode_id(stored.id) ), "Rename credentials", submit_text="Rename" ) \ - .add_text( "new_name", "Credentials Name", value=stored.name ) + return trans.show_form( + web.FormBuilder( url_for( id=trans.security.encode_id(stored.id) ), "Rename credentials", submit_text="Rename" ) + .add_text( "new_name", "Credentials Name", value=stored.name ) ) @web.expose @web.require_login( "use Galaxy cloud" ) @@ -473,8 +403,9 @@ trans.set_message( "Instance renamed to '%s'." % new_name ) return self.list( trans ) else: - return form( url_for( id=trans.security.encode_id(instance.id) ), "Rename instance", submit_text="Rename" ) \ - .add_text( "new_name", "Instance name", value=instance.name ) + return trans.show_form( + web.FormBuilder( url_for( id=trans.security.encode_id(instance.id) ), "Rename instance", submit_text="Rename" ) + .add_text( "new_name", "Instance name", value=instance.name ) ) @web.expose @web.require_login( "add credentials" ) @@ -485,14 +416,14 @@ user = trans.get_user() error = {} - if credName: - if len( credName ) > 255: - error['cred_error'] = "Credentials name exceeds maximum allowable length." - elif trans.app.model.CloudUserCredentials.filter( + if credName or providerName or accessKey or secretKey: + if credName=='' or len( credName ) > 255: + error['cred_error'] = "Credentials name must be between 1 and 255 characters in length." + elif trans.app.model.CloudUserCredentials.filter_by( user=user ).filter( trans.app.model.CloudUserCredentials.table.c.name==credName ).first(): error['cred_error'] = "Credentials with that name already exist." - elif ( ( providerName.lower()!='ec2' ) and ( providerName.lower()!='eucalyptus' ) ): - error['provider_error'] = "You specified an unsupported cloud provider." + elif providerName=='': + error['provider_error'] = "You must select cloud provider associated with these credentials." elif accessKey=='' or len( accessKey ) > 255: error['access_key_error'] = "Access key must be between 1 and 255 characters long." elif secretKey=='' or len( secretKey ) > 255: @@ -504,7 +435,8 @@ credentials.user = user credentials.access_key = accessKey credentials.secret_key = secretKey - credentials.provider_name = providerName.lower() + provider = get_provider( trans, providerName ) + credentials.provider = provider # Persist session = trans.sa_session session.save_or_update( credentials ) @@ -516,12 +448,14 @@ # self.makeDefault( trans, credentials.id) return self.list( trans ) - return trans.fill_template( "cloud/add_credentials.mako", \ - credName = credName, \ - providerName = providerName, \ - accessKey = accessKey, \ - secretKey = secretKey, \ - error = error + providers = trans.sa_session.query( model.CloudProvider ).filter_by( user=user ).all() + return trans.fill_template( "cloud/add_credentials.mako", + credName = credName, + providerName = providerName, + accessKey = accessKey, + secretKey = secretKey, + error = error, + providers = providers ) # return trans.show_form( @@ -544,6 +478,13 @@ credDetails = stored ) @web.expose + @web.require_login( "test cloud credentials" ) + def test_cred( self, trans, id=None ): + """ + Tests credentials provided by user with selected cloud provider + """ + + @web.expose @web.require_login( "view instance details" ) def viewInstance( self, trans, id=None ): """ @@ -574,6 +515,97 @@ return self.list( trans ) @web.expose + @web.require_login( "add provider" ) + def add_provider( self, trans, name='', type='', region_name='', region_endpoint='', is_secure='', host='', port='', proxy='', proxy_port='', + proxy_user='', proxy_pass='', debug='', https_connection_factory='', path='' ): + user = trans.get_user() + error = {} + + # Check if Amazon EC2 has already been registered by this user + ec2_registered = trans.sa_session.query( model.CloudProvider ).filter_by( user=user, type='ec2' ).first() + + if region_name or region_endpoint or name or is_secure or port or proxy or debug or path: + if trans.app.model.CloudProvider \ + .filter_by (user=user, name=name) \ + .first(): + error['name_error'] = "A provider with that name already exist." + elif name=='' or len( name ) > 255: + error['name_error'] = "Provider name must be between 1 and 255 characters long." + elif type=='': + error['type_error'] = "Provider type must be selected." + elif ec2_registered: + error['type_error'] = "Amazon EC2 has already been registered as a provider." + elif is_secure!=0 or is_secure!=1: + error['is_secure_error'] = "Field 'is secure' can only take on a value '0' or '1'." + else: + provider = model.CloudProvider() + provider.user = user + provider.type = type + provider.name = name + provider.region_name = region_name + provider.region_endpoint = region_endpoint + if is_secure=='0': + provider.is_secure = False + else: + provider.is_secure = True + provider.host = host + provider.port = port + provider.proxy = proxy + provider.proxy_port = proxy_port + provider.proxy_user = proxy_user + provider.proxy_pass = proxy_pass + provider.debug = debug + provider.https_connection_factory = https_connection_factory + provider.path = path + # Persist + session = trans.sa_session + session.save_or_update( provider ) + session.flush() + # Log and display the management page + trans.log_event( "User configured new cloud provider: '%s'" % name ) + trans.set_message( "New cloud provider '%s' added." % name ) + return self.list( trans ) + + return trans.fill_template( "cloud/add_provider.mako", + name = name, + type = type, + region_name = region_name, + region_endpoint = region_endpoint, + is_secure = is_secure, + host = host, + port = port, + proxy = proxy, + proxy_port = proxy_port, + proxy_user = proxy_user, + proxy_pass = proxy_pass, + debug = debug, + https_connection_factory = https_connection_factory, + path = path, + error = error + ) + + @web.expose + @web.require_login( "add Amazon EC2 provider" ) + def add_ec2( self, trans ): + """ Default provider setup for Amazon's EC2. """ + user = trans.get_user() + # Check if EC2 has already been registered by this user. + exists = trans.sa_session.query( model.CloudProvider ) \ + .filter_by( user=user, type='ec2' ).first() + + if not exists: + provider = model.CloudProvider() + provider.user = user + provider.type = 'ec2' + provider.name = 'EC2' + # Persist + session = trans.sa_session + session.save_or_update( provider ) + session.flush() + + return self.list( trans ) + + @web.expose @web.require_login( "edit workflows" ) def editor( self, trans, id=None ): """ @@ -1012,25 +1044,32 @@ workflows=workflows, shared_by_others=shared_by_others, ids_in_menu=ids_in_menu ) + @web.json + def json_update( self, trans ): + user = trans.get_user() + UCIs = trans.sa_session.query( model.UCI ).filter_by( user=user ).filter( model.UCI.c.state != uci_states.DELETED ).all() + insd = {} # instance name-state dict + for uci in UCIs: + dict = {} + dict['id'] = uci.id + dict['state'] = uci.state + if uci.launch_time != None: + dict['launch_time'] = str(uci.launch_time) + dict['time_ago'] = str(date.distance_of_time_in_words(uci.launch_time, date.datetime.utcnow() ) ) + else: + dict['launch_time'] = None + dict['time_ago'] = None + insd[uci.name] = dict + return insd ## ---- Utility methods ------------------------------------------------------- -def get_UCIs_state( trans ): +def get_provider( trans, name ): user = trans.get_user() - instances = trans.sa_session.query( model.UCI ).filter_by( user=user ).filter( model.UCI.c.state != uci_states.DELETED ).all() - insd = {} # instance name-state dict - for inst in instances: - insd[inst.name] = inst.state + return trans.app.model.CloudProvider \ + .filter_by (user=user, name=name) \ + .first() - -def get_UCIs_time_ago( trans ): - user = trans.get_user() - instances = trans.sa_session.query( model.UCI ).filter_by( user=user ).all() - intad = {} # instance name-time-ago dict - for inst in instances: - if inst.launch_time != None: - intad[inst.name] = str(date.distance_of_time_in_words (inst.launch_time, date.datetime.utcnow() ) ) - def get_stored_credentials( trans, id, check_ownership=True ): """ Get StoredUserCredentials from the database by id, verifying ownership. @@ -1085,14 +1124,18 @@ # Looks good return live -def get_mi( trans, size='m1.small' ): +def get_mi( trans, uci, size='m1.small' ): """ Get appropriate machine image (mi) based on instance size. TODO: Dummy method - need to implement logic For valid sizes, see http://aws.amazon.com/ec2/instance-types/ """ - return trans.app.model.CloudImage.filter( - trans.app.model.CloudImage.table.c.id==2).first() + if uci.credentials.provider.type == 'ec2': + return trans.app.model.CloudImage.filter( + trans.app.model.CloudImage.table.c.id==2).first() + else: + return trans.app.model.CloudImage.filter( + trans.app.model.CloudImage.table.c.id==1).first() def get_stores( trans, uci ): """ diff -r 7f3054622e87 -r 676fae1a35f9 templates/cloud/add_credentials.mako --- a/templates/cloud/add_credentials.mako Wed Oct 21 11:40:51 2009 -0400 +++ b/templates/cloud/add_credentials.mako Fri Oct 23 17:41:35 2009 -0400 @@ -15,11 +15,13 @@ ${header} %endif +%if providers: + <div class="form"> <div class="form-title">Add credentials</div> <div class="form-body"> <form name="Add credentials" action="${h.url_for( action='add' )}" method="post" > - + <% cls = "form-row" if error.has_key('cred_error'): @@ -45,11 +47,15 @@ <div class="${cls}"> <label>Cloud provider name:</label> <div class="form-row-input"> - <select name="providerName" style="width:40em"> - <option value="ec2">Amazon EC2</option> - <option value="eucalyptus">Eucalpytus Public Cloud (EPC)</option> - </select> - </div> + <select name="providerName" style="width:40em"> + <option value="">Select Provider...</option> + %for provider in providers: + <option value="${provider.name}">${provider.name}</option> + %endfor + </select> + <br/>or <a href="${h.url_for( action='add_provider' )}"> + <span>register additional cloud provider</span></a> + </div> %if error.has_key('provider_error'): <div class="form-row-error-message">${error['provider_error']}</div> %endif @@ -95,3 +101,12 @@ </form> </div> </div> + +%else: + In order to add credentials, desired cloud provider needs to be registered.<p/> + Register <a href="${h.url_for( action='add_ec2' )}"> + <span>Amazon EC2 automatically</span></a> + or add + <a href="${h.url_for( action='add_provider' )}"> + <span>custom cloud provider</span></a>. +%endif diff -r 7f3054622e87 -r 676fae1a35f9 templates/cloud/add_provider.mako --- /dev/null Thu Jan 01 00:00:00 1970 +0000 +++ b/templates/cloud/add_provider.mako Fri Oct 23 17:41:35 2009 -0400 @@ -0,0 +1,225 @@ +<% _=n_ %> +<%inherit file="/base.mako"/> +<%def name="title()">Add provider</%def> + +<%def name="javascripts()"> +${parent.javascripts()} +<script type="text/javascript"> +$(function(){ + + $("input:text:first").focus(); + + $("#type").change(function() { + if ($(this).val() == 'ec2') { + $("#name").val("EC2"); + } + }); +}) + + +function af(){ + + if ( $("#autofill").attr('checked') ) { + $("#region_name").val("eucalyptus"); + $("#region_endpoint").val("mayhem9.cs.ucsb.edu"); + $("#is_secure").val("0"); + $("#port").val("8773"); + $("#path").val("/services/Eucalyptus"); + } + else { + $("#region_name").val(""); + $("#region_endpoint").val(""); + $("#is_secure").val(""); + $("#port").val(""); + $("#path").val(""); + } + +} + +</script> +</%def> + +%if header: + ${header} +%endif + +<div class="form"> + <div class="form-title">Add cloud provider</div> + <div class="form-body"> + <form name="add_provider_form" action="${h.url_for( action='add_provider' )}" method="post" > + <% + cls = "form-row" + if error.has_key('type_error'): + cls += " form-row-error" + %> + <div class="${cls}"> + <label>Provider type:</label> + <div class="form-row-input"> + <select id="type" name="type" style="width:40em"> + <option value="">Select Provider...</option> + <option value="eucalyptus">Eucalyptus</option> + <option value="ec2">Amazon EC2</option> + </select> + <br/><input type="checkbox" id="autofill" onclick="javascript:af()"> + auto fill using Eucalyptus Public Cloud values + </div> + %if error.has_key('type_error'): + <div class="form-row-error-message">${error['type_error']}</div> + %endif + <div style="clear: both"></div> + </div> + + <% + cls = "form-row" + if error.has_key('name_error'): + cls += " form-row-error" + %> + <div class="${cls}"> + <label>Provider name:</label> + <div class="form-row-input"> + <input type="text" id="name" name="name" value="${name}" size="40"> + </div> + %if error.has_key('name_error'): + <div class="form-row-error-message">${error['name_error']}</div> + %endif + <div style="clear: both"></div> + </div> + + <% + cls = "form-row" + %> + <div class="${cls}"> + <label>Region name:</label> + <div class="form-row-input"> + <input type="text" name="region_name" id="region_name" value="${region_name}" size="40"> + </div> + <div style="clear: both"></div> + </div> + + <% + cls = "form-row" + %> + <div class="${cls}"> + <label>Region endpoint:</label> + <div class="form-row-input"> + <input type="text" name="region_endpoint" id="region_endpoint" value="${region_endpoint}" size="40"> + </div> + <div style="clear: both"></div> + </div> + + <% + cls = "form-row" + %> + <div class="${cls}"> + <label>Is secure ('O' for False or '1' for True):</label> + <div class="form-row-input"> + <input type="text" name="is_secure" id="is_secure" value="${is_secure}" size="40"> + </div> + <div style="clear: both"></div> + </div> + + <% + cls = "form-row" + %> + <div class="${cls}"> + <label>Host:</label> + <div class="form-row-input"> + <input type="text" name="host" value="${host}" size="40"> + </div> + <div style="clear: both"></div> + </div> + + <% + cls = "form-row" + %> + <div class="${cls}"> + <label>Port:</label> + <div class="form-row-input"> + <input type="text" name="port" id="port" value="${port}" size="40"> + </div> + <div style="clear: both"></div> + </div> + + <% + cls = "form-row" + %> + <div class="${cls}"> + <label>Proxy:</label> + <div class="form-row-input"> + <input type="text" name="proxy" value="${proxy}" size="40"> + </div> + <div style="clear: both"></div> + </div> + + <% + cls = "form-row" + %> + <div class="${cls}"> + <label>Proxy port:</label> + <div class="form-row-input"> + <input type="text" name="proxy_port" value="${proxy_port}" size="40"> + </div> + <div style="clear: both"></div> + </div> + + <% + cls = "form-row" + %> + <div class="${cls}"> + <label>Proxy user:</label> + <div class="form-row-input"> + <input type="text" name="proxy_user" value="${proxy_user}" size="40"> + </div> + <div style="clear: both"></div> + </div> + + <% + cls = "form-row" + %> + <div class="${cls}"> + <label>Proxy pass:</label> + <div class="form-row-input"> + <input type="text" name="proxy_pass" value="${proxy_pass}" size="40"> + </div> + <div style="clear: both"></div> + </div> + + <% + cls = "form-row" + %> + <div class="${cls}"> + <label>Debug:</label> + <div class="form-row-input"> + <input type="text" name="debug" value="${debug}" size="40"> + </div> + <div style="clear: both"></div> + </div> + + <% + cls = "form-row" + %> + <div class="${cls}"> + <label>HTTPS connection factory:</label> + <div class="form-row-input"> + <input type="text" name="https_connection_factory" value="${https_connection_factory}" size="40"> + </div> + <div style="clear: both"></div> + </div> + + <% + cls = "form-row" + %> + <div class="${cls}"> + <label>Path:</label> + <div class="form-row-input"> + <input type="text" name="path" id="path" value="${path}" size="40"> + </div> + <div style="clear: both"></div> + </div> + + <div class="form-row"><input type="submit" value="Add"></div> + + </form> + + </div> +</div> diff -r 7f3054622e87 -r 676fae1a35f9 templates/cloud/configure_cloud.mako --- a/templates/cloud/configure_cloud.mako Wed Oct 21 11:40:51 2009 -0400 +++ b/templates/cloud/configure_cloud.mako Fri Oct 23 17:41:35 2009 -0400 @@ -10,14 +10,38 @@ messagetype = "done" %> - - -<p /> <div class="${messagetype}message"> ${message} </div> %endif +<%def name="javascripts()"> +${parent.javascripts()} +${h.js( "jquery" )} + +<script type="text/javascript"> + function update_state() { + $.getJSON( "${h.url_for( action='json_update' )}", {}, function ( data ) { + for (var i in data) { + var elem = '#' + data[i].id; + $(elem + "-state").text( data[i].state ); + if (data[i].launch_time) { + $(elem + "-launch_time").text( data[i].launch_time.substring(0, 16 ) + " (" + data[i].time_ago + ")" ); + } + else { + $(elem + "-launch_time").text( "N/A" ); + } + } + }); + setTimeout("update_state()", 10000); + } + + $(function() { + update_state(); + }); +</script> +</%def> + <h2>Galaxy in the clouds</h2> %if cloudCredentials: @@ -34,22 +58,23 @@ <table class="mange-table colored" border="0" cellspacing="0" cellpadding="0" width="100%"> <tr class="header"> <th>Credentials name</th> - <th>Provider</th> + <th>Provider name (type)</th> <th></th> </tr> %for i, cloudCredential in enumerate( cloudCredentials ): <tr> <td> ${cloudCredential.name} - <a id="cr-${i}-popup" class="popup-arrow" style="display: none;">▼</a> + <a id="cr-${i}-popup" class="popup-arrow" style="display: none;">▼</a> </td> - <td>${cloudCredential.provider_name}</td> - + <td> + ${cloudCredential.provider.name} + (${cloudCredential.provider.type}) + </td> <td> <div popupmenu="cr-${i}-popup"> <a class="action-button" href="${h.url_for( action='view', id=trans.security.encode_id(cloudCredential.id) )}">View</a> - <a class="action-button" href="${h.url_for( action='rename', id=trans.security.encode_id(cloudCredential.id) )}">Rename</a> - <a class="action-button" href="${h.url_for( action='makeDefault', id=trans.security.encode_id(cloudCredential.id) )}" target="_parent">Make default</a> + <a class="action-button" href="${h.url_for( action='rename', id=trans.security.encode_id(cloudCredential.id) )}">Rename</a> <a class="action-button" confirm="Are you sure you want to delete credentials '${cloudCredential.name}'?" href="${h.url_for( action='delete', id=trans.security.encode_id(cloudCredential.id) )}">Delete</a> </div> </td> @@ -60,7 +85,7 @@ ## ***************************************************** ## Manage live instances <p /> - <h2>Manage cloud instances</h2> + <h2>Manage your cloud instances</h2> <ul class="manage-table-actions"> <li> <a class="action-button" href="${h.url_for( action='configureNew' )}"> @@ -77,7 +102,7 @@ <colgroup width="25%"></colgroup> <colgroup width="10%"></colgroup> <tr class="header"> - <th>Your instances</th> + <th>Instance name (credentials)</th> <th>Storage size (GB)</th> <th>State</th> <th>Alive since</th> @@ -88,12 +113,12 @@ %for i, liveInstance in enumerate( liveInstances ): <tr> <td> - ${liveInstance.name} + ${liveInstance.name} (${liveInstance.credentials.name}) <a id="li-${i}-popup" class="popup-arrow" style="display: none;">▼</a> </td> <td>${str(liveInstance.total_size)}</td> - <td>${str(liveInstance.state)}</td> - <td> + <td id="${ liveInstance.id }-state">${str(liveInstance.state)}</td> + <td id="${ liveInstance.id }-launch_time"> ##${str(liveInstance.launch_time)[:16]} <% #from datetime import datetime @@ -120,9 +145,18 @@ %for j, instance in enumerate( liveInstance.instance ): ## TODO: Once more instances will be running under the same liveInstance, additional logic will need to be added to account for that %if instance.state == "running": - <a class="action-button" href="http://${instance.public_dns}" target="_blank"> - <span>Access Galaxy</span> - <img src="${h.url_for('/static/images/silk/resultset_next.png')}" /></a></div> + ## Wait until Galaxy server starts to show 'Access Galaxy' button + <% + import urllib2 + try: + urllib2.urlopen("http://"+instance.public_dns) + context.write( '<a class="action-button" href="http://'+instance.public_dns+'" target="_blank">' ) + context.write( '<span>Access Galaxy</span>' ) + context.write( '<img src="'+h.url_for('/static/images/silk/resultset_next.png')+'" /></a></div>' ) + except urllib2.URLError: + context.write( '<span>Galaxy starting...</span>' ) + %> + %endif %endfor </td> @@ -164,17 +198,39 @@ %for i, prevInstance in enumerate( prevInstances ): <tr> <td> - ${prevInstance.name} + ${prevInstance.name} (${prevInstance.credentials.name}) <a id="pi-${i}-popup" class="popup-arrow" style="display: none;">▼</a> </td> <td>${str(prevInstance.total_size)}</td>  - <td>${str(prevInstance.state)}</td> + <td> + <%state = str(prevInstance.state)%> + %if state =='error': + <div id="short"> + <a onclick="document.getElementById('full').style.display = 'block'; + document.getElementById('short').style.display = 'none'; return 0" + href="javascript:void(0)"> + error + </a> + </div> + <div id="full" style="DISPLAY: none"> + <a onclick="document.getElementById('short').style.display = 'block'; + document.getElementById('full').style.display = 'none'; return 0;" + href="javascript:void(0)"> + ${str(prevInstance.error)} + </a> + </div> + %else: + ${str(prevInstance.state)} + %endif + </td> <td>N/A</td> <td> <div popupmenu="pi-${i}-popup"> - <a class="action-button" href="${h.url_for( action='start', id=trans.security.encode_id(prevInstance.id) )}">Start</a> - <a class="action-button" href="${h.url_for( action='renameInstance', id=trans.security.encode_id(prevInstance.id) )}">Rename</a> + <a class="action-button" href="${h.url_for( action='start', id=trans.security.encode_id(prevInstance.id), type='m1.small' )}"> Start m1.small</a> + <a class="action-button" href="${h.url_for( action='start', id=trans.security.encode_id(prevInstance.id), type='c1.medium' )}"> Start c1.medium</a> + <a class="action-button" href="${h.url_for( action='renameInstance', id=trans.security.encode_id(prevInstance.id) )}">Rename</a> <a class="action-button" href="${h.url_for( action='addStorage', id=trans.security.encode_id(prevInstance.id) )}" target="_parent">Add storage</a> + <a class="action-button" href="${h.url_for( action='usageReport' )}">Usage report</a> <a class="action-button" confirm="Are you sure you want to delete instance '${prevInstance.name}'? This will delete all of your data assocaiated with this instance!" href="${h.url_for( action='deleteInstance', id=trans.security.encode_id(prevInstance.id) )}">Delete</a> </div> </td> diff -r 7f3054622e87 -r 676fae1a35f9 templates/cloud/view.mako --- a/templates/cloud/view.mako Wed Oct 21 11:40:51 2009 -0400 +++ b/templates/cloud/view.mako Fri Oct 23 17:41:35 2009 -0400 @@ -38,8 +38,12 @@ </td> </tr> <tr> - <td> Cloud provider: </td> - <td> ${str(credDetails.provider_name)[:16]}</td> + <td> Cloud provider type: </td> + <td> ${str(credDetails.provider.type)[:16]}</td> + </tr> + <tr> + <td> Cloud provider name: </td> + <td> ${str(credDetails.provider.name)[:16]}</td> </tr> <tr> <td> Access key: </td> @@ -67,6 +71,79 @@ </div> </td> </tr> + <tr><td><b>Additional cloud provider information (if available):</b></td></tr> + %if credDetails.provider.region_connection != None: + <tr> + <td> Region connection: </td> + <td> ${credDetails.provider.region_connection} </td> + </tr> + %endif + %if credDetails.provider.region_name != None: + <tr> + <td> Region name: </td> + <td> ${credDetails.provider.region_name} </td> + </tr> + %endif + %if credDetails.provider.region_endpoint != None: + <tr> + <td> Region endpoint: </td> + <td> ${credDetails.provider.region_endpoint} </td> + </tr> + %endif + %if credDetails.provider.is_secure != None: + <tr> + <td> Is secure: </td> + <td> ${credDetails.provider.is_secure} </td> + </tr> + %endif + %if credDetails.provider.host != None: + <tr> + <td> Host: </td> + <td> ${credDetails.provider.host} </td> + </tr> + %endif + %if credDetails.provider.port != None: + <tr> + <td> Port: </td> + <td> ${credDetails.provider.port} </td> + </tr> + %endif + %if credDetails.provider.proxy != None: + <tr> + <td> Proxy: </td> + <td> ${credDetails.provider.proxy} </td> + </tr> + %endif + %if credDetails.provider.proxy_port != None: + <tr> + <td> Proxy port: </td> + <td> ${credDetails.provider.proxy_port} </td> + </tr> + %endif + %if credDetails.provider.proxy_pass != None: + <tr> + <td> Proxy pass: </td> + <td> ${credDetails.provider.proxy_pass} </td> + </tr> + %endif + %if credDetails.provider.debug != None: + <tr> + <td> Debug: </td> + <td> ${credDetails.provider.debug} </td> + </tr> + %endif + %if credDetails.provider.https_connection_factory != None: + <tr> + <td> HTTPS connection factory: </td> + <td> ${credDetails.provider.https_connection_factory} </td> + </tr> + %endif + %if credDetails.provider.path != None: + <tr> + <td> Path: </td> + <td> ${credDetails.provider.path} </td> + </tr> + %endif </table> %else: There are no credentials under that name. diff -r 7f3054622e87 -r 676fae1a35f9 templates/cloud/view_usage.mako --- /dev/null Thu Jan 01 00:00:00 1970 +0000 +++ b/templates/cloud/view_usage.mako Fri Oct 23 17:41:35 2009 -0400 @@ -0,0 +1,47 @@ +<%inherit file="/base.mako"/> + +<%def name="title()">Cloud home</%def> + +%if message: +<% + try: + messagetype + except: + messagetype = "done" +%> + + + +<p /> +<div class="${messagetype}message"> + ${message} +</div> +%endif + +<h2>Instance usage report</h2> + +%if prevInstances: + <table class="mange-table colored" border="0" cellspacing="0" cellpadding="0" width="100%"> + <colgroup width="2%"></colgroup> + <colgroup width="20%"></colgroup> + <colgroup width="20%"></colgroup> + <colgroup width="5%"></colgroup> + <tr class="header"> + <th>#</th> + <th>Launch time</th> + <th>Termination time</th> + <th>Type</th> + <th></th> + </tr> + %for i, prevInstance in enumerate( prevInstances ): + <tr> + <td>${i+1}</td> + <td>${str(prevInstance.launch_time)[:16]} UCT</td> + <td>${str(prevInstance.stop_time)[:16]} UCT</td> + <td>${prevInstance.type}</td> + </tr> + %endfor + </table> +%else: + Selected instance has no record of being used. +%endif \ No newline at end of file

1 0

[hg] galaxy 3068: Decoupled cloud controller from cloud provider...
by Greg Von Kuster 23 Nov '09

23 Nov '09

details: http://www.bx.psu.edu/hg/galaxy/rev/6aab50510e43 changeset: 3068:6aab50510e43 user: Enis Afgan <afgane(a)gmail.com> date: Wed Oct 14 19:20:11 2009 -0400 description: Decoupled cloud controller from cloud providers. Added cloud manager that runs as a daemon and pools database for change of ststus (i.e., user action). Implemented code to support 'eucalyptus' provider (tested with Eucalyptus Public Cloud). diffstat: lib/galaxy/cloud/__init__.py | 367 ++++++++++++++++- lib/galaxy/cloud/providers/eucalyptus.py | 476 +++++++++++++++++----- lib/galaxy/model/mapping.py | 2 +- lib/galaxy/model/migrate/versions/0014_cloud_tables.py | 8 +- lib/galaxy/web/controllers/cloud.py | 137 ++++-- templates/cloud/add_credentials.mako | 97 ++++ 6 files changed, 895 insertions(+), 192 deletions(-) diffs (1401 lines): diff -r 7c438fd3cf4a -r 6aab50510e43 lib/galaxy/cloud/__init__.py --- a/lib/galaxy/cloud/__init__.py Fri Oct 02 18:31:32 2009 -0400 +++ b/lib/galaxy/cloud/__init__.py Wed Oct 14 19:20:11 2009 -0400 @@ -7,6 +7,7 @@ from galaxy.datatypes.interval import * from galaxy.datatypes import metadata from galaxy.util.bunch import Bunch +from sqlalchemy import or_ import pkg_resources pkg_resources.require( "PasteDeploy" ) @@ -147,13 +148,18 @@ # HACK: Delay until after forking, we need a way to do post fork notification!!! time.sleep( 10 ) + cnt = 0 # Run global update only periodically so keep counter variable while self.running: try: # log.debug( "Calling monitor_step" ) self.__monitor_step() + if cnt%30 == 0: # Run global update every 30 seconds + self.provider.update() + cnt = 0 except: log.exception( "Exception in cloud manager monitor_step" ) # Sleep + cnt += 1 self.sleeper.sleep( 2 ) def __monitor_step( self ): @@ -167,31 +173,53 @@ it is marked as having errors and removed from the queue. Otherwise, the job is dispatched. """ - # Get an orm session + # Get an orm (object relational mapping) session session = mapping.Session() # Pull all new jobs from the queue at once - new_jobs = [] - new_instances = [] - new_UCIs = [] - stop_UCIs = [] + new_requests = [] +# new_instances = [] +# new_UCIs = [] +# stop_UCIs = [] +# delete_UCIs = [] # for r in session.query( model.cloud_instance ).filter( model.cloud_instance.s.state == model.cloud_instance.states.NEW ).all(): # new_instances - for r in session.query( model.UCI ).filter( model.UCI.c.state == "new" ).all(): - new_UCIs.append( r ) - for r in new_UCIs: - self.provider.createUCI( r ) + for r in session.query( model.UCI ) \ + .filter( or_( model.UCI.c.state=="newUCI", + model.UCI.c.state=="submittedUCI", + model.UCI.c.state=="shutting-downUCI", + model.UCI.c.state=="deletingUCI" ) ) \ + .all(): + uci = UCIwrapper( r ) + new_requests.append( uci ) +# log.debug( 'new_requests: %s' % new_requests ) + for uci in new_requests: + session.clear() +# log.debug( 'r.name: %s, state: %s' % ( r.name, r.state ) ) +# session.save_or_update( r ) +# session.flush() + self.provider.put( uci ) - for r in session.query( model.UCI ).filter( model.UCI.c.state == "submitted" ).all(): - new_instances.append( r ) - for r in new_instances: - self.provider.startUCI( r ) + # Done with the session + mapping.Session.remove() - for r in session.query( model.UCI ).filter( model.UCI.c.state == "terminating" ).all(): - stop_UCIs.append( r ) - for r in stop_UCIs: - self.provider.stopUCI( r ) +# for r in session.query( model.UCI ).filter( model.UCI.c.state == "submitted" ).all(): +# new_instances.append( r ) +# for r in new_instances: +# self.provider.startUCI( r ) +# +# for r in session.query( model.UCI ).filter( model.UCI.c.state == "shutting-down" ).all(): +# stop_UCIs.append( r ) +# for r in stop_UCIs: +# self.provider.stopUCI( r ) +# +# for r in session.query( model.UCI ).filter( model.UCI.c.state == "deleting" ).all(): +# delete_UCIs.append( r ) +# for r in delete_UCIs: +# self.provider.deleteUCI( r ) + + # if self.track_jobs_in_database: # for j in session.query( model.Job ).options( lazyload( "external_output_metadata" ), lazyload( "parameters" ) ).filter( model.Job.c.state == model.Job.states.NEW ).all(): @@ -327,6 +355,281 @@ log.info( "cloud manager stopped" ) self.dispatcher.shutdown() +class UCIwrapper( object ): + """ + Wraps 'model.UCI' with convenience methods for state management + """ + def __init__( self, uci ): + self.uci_id = uci.id + + # --------- Setter methods ----------------- + + def change_state( self, uci_state=None, instance_id=None, i_state=None ): + """ + Sets state for UCI and/or UCI's instance with instance_id as provided by cloud provider and stored in local + Galaxy database. + Need to provide either state for the UCI or instance_id and it's state or all arguments. + """ +# log.debug( "Changing state - new uci_state: %s, instance_id: %s, i_state: %s" % ( uci_state, instance_id, i_state ) ) + if uci_state is not None: + uci = model.UCI.get( self.uci_id ) + uci.refresh() + uci.state = uci_state + uci.flush() + if ( instance_id is not None ) and ( i_state is not None ): + instance = model.CloudInstance.filter_by( uci_id=self.uci_id, instance_id=instance_id).first() + instance.state = i_state + instance.flush() + + def set_mi( self, i_index, mi_id ): + """ + Sets Machine Image (MI), e.g., 'ami-66fa190f', for UCI's instance with given index as it + is stored in local Galaxy database. + """ + mi = model.CloudImage.filter( model.CloudImage.c.image_id==mi_id ).first() + instance = model.CloudInstance.get( i_index ) + instance.image = mi + instance.flush() + + def set_key_pair( self, i_index, key_name, key_material=None ): + """ + Single UCI may instantiate many instances, i_index refers to the numeric index + of instance controlled by this UCI as it is stored in local DB (see get_instances_ids()). + """ + instance = model.CloudInstance.get( i_index ) + instance.keypair_name = key_name + if key_material is not None: + instance.keypair_material = key_material + instance.flush() + + def set_launch_time( self, launch_time, i_index=None, i_id=None ): + """ + Stores launch time in local database for instance with specified index (as it is stored in local + Galaxy database) or with specified instance ID (as obtained from the cloud provider AND stored + in local Galaxy Database). Only one of i_index or i_id needs to be provided. + """ + if i_index != None: + instance = model.CloudInstance.get( i_index ) + instance.launch_time = launch_time + instance.flush() + elif i_id != None: + instance = model.CloudInstance.filter_by( uci_id=self.uci_id, instance_id=i_id).first() + instance.launch_time = launch_time + instance.flush() + + def set_uci_launch_time( self, launch_time ): + uci = model.UCI.get( self.uci_id ) + uci.refresh() + uci.launch_time = launch_time + uci.flush() + + def set_stop_time( self, stop_time, i_index=None, i_id=None ): + if i_index != None: + instance = model.CloudInstance.get( i_index ) + instance.stop_time = stop_time + instance.flush() + elif i_id != None: + instance = model.CloudInstance.filter_by( uci_id=self.uci_id, instance_id=i_id).first() + instance.stop_time = stop_time + instance.flush() + + def reset_uci_launch_time( self ): + uci = model.UCI.get( self.uci_id ) + uci.refresh() + uci.launch_time = None + uci.flush() + + def set_reservation_id( self, i_index, reservation_id ): + instance = model.CloudInstance.get( i_index ) + instance.reservation_id = reservation_id + instance.flush() + + def set_instance_id( self, i_index, instance_id ): + """ + i_index refers to UCI's instance ID as stored in local database + instance_id refers to real-world, cloud resource ID (e.g., 'i-78hd823a') + """ + instance = model.CloudInstance.get( i_index ) + instance.instance_id = instance_id + instance.flush() + + def set_public_dns( self, instance_id, public_dns ): + uci = model.UCI.get( self.uci_id ) + uci.refresh() + uci.instance[instance_id].public_dns = public_dns + uci.instance[instance_id].flush() + + def set_private_dns( self, instance_id, private_dns ): + uci = model.UCI.get( self.uci_id ) + uci.refresh() + uci.instance[instance_id].private_dns = private_dns + uci.instance[instance_id].flush() + + def set_store_device( self, store_id, device ): + uci = model.UCI.get( self.uci_id ) + uci.refresh() + uci.store[store_id].device = device + uci.store[store_id].flush() + + def set_store_status( self, store_id, status ): + uci = model.UCI.get( self.uci_id ) + uci.refresh() + uci.store[store_id].status = status + uci.store[store_id].flush() + + def set_store_availability_zone( self, store_id, availability_zone ): + uci = model.UCI.get( self.uci_id ) + uci.refresh() + uci.store[store_id].availability_zone = availability_zone + uci.store[store_id].flush() + + def set_store_volume_id( self, store_id, volume_id ): + """ + Given store ID associated with this UCI, set volume ID as it is registered + on the cloud provider (e.g., vol-39890501) + """ + uci = model.UCI.get( self.uci_id ) + uci.refresh() + uci.store[store_id].volume_id = volume_id + uci.store[store_id].flush() + + def set_store_instance( self, store_id, instance_id ): + """ Stores instance ID that given store volume is attached to. """ + uci = model.UCI.get( self.uci_id ) + uci.refresh() + uci.store[store_id].i_id = instance_id + uci.store[store_id].flush() + + # --------- Getter methods ----------------- + + def get_instances_indexes( self, state=None ): + """ + Returns indexes of instances associated with given UCI as they are stored in local Galaxy database and + whose state corresponds to passed argument. Returned values enable indexing instances from local Galaxy database. + """ + uci = model.UCI.get( self.uci_id ) + uci.refresh() + instances = model.CloudInstance.filter_by( uci=uci ).filter( model.CloudInstance.c.state==state ).all() + il = [] + for i in instances: + il.append( i.id ) + + return il + + def get_type( self, i_index ): + instance = model.CloudInstance.get( i_index ) + return instance.type + + def get_state( self ): + uci = model.UCI.get( self.uci_id ) + uci.refresh() + return uci.state + + def get_instance_state( self, instance_id ): + uci = model.UCI.get( self.uci_id ) + uci.refresh() + return uci.instance[instance_id].state + + def get_instances_ids( self ): + """ + Returns list IDs of all instances' associated with this UCI that are not in 'terminated' state + (e.g., ['i-402906D2', 'i-q0290dsD2'] ). + """ + il = model.CloudInstance.filter_by( uci_id=self.uci_id ).filter( model.CloudInstance.c.state != 'terminated' ).all() + instanceList = [] + for i in il: + instanceList.append( i.instance_id ) + return instanceList + + def get_name( self ): + uci = model.UCI.get( self.uci_id ) + uci.refresh() + return uci.name + + def get_key_pair_name( self, i_index=None, i_id=None ): + """ + Given EITHER instance index as it is stored in local Galaxy database OR instance ID as it is + obtained from cloud provider and stored in local Galaxy database, return keypair name assocaited + with given instance. + """ + if i_index != None: + instance = model.CloudInstance.get( i_index ) + return instance.keypair_name + elif i_id != None: + instance = model.CloudInstance.filter_by( uci_id=self.uci_id, instance_id=i_id).first() + return instance.keypair_name + + def get_access_key( self ): + uci = model.UCI.get( self.uci_id ) + uci.refresh() + return uci.credentials.access_key + + def get_secret_key( self ): + uci = model.UCI.get( self.uci_id ) + uci.refresh() + return uci.credentials.secret_key + + def get_mi_id( self, instance_id=0 ): + uci = model.UCI.get( self.uci_id ) + uci.refresh() + return uci.instance[instance_id].mi_id + + def get_public_dns( self, instance_id=0 ): + uci = model.UCI.get( self.uci_id ) + uci.refresh() + return uci.instance[instance_id].public_dns + + def get_private_dns( self, instance_id=0 ): + uci = model.UCI.get( self.uci_id ) + uci.refresh() + return uci.instance[instance_id].private_dns + + def get_store_availability_zone( self, store_id ): + uci = model.UCI.get( self.uci_id ) + uci.refresh() + return uci.store[store_id].availability_zone + + def get_store_size( self, store_id ): + uci = model.UCI.get( self.uci_id ) + uci.refresh() + return uci.store[store_id].size + + def get_store_volume_id( self, store_id ): + """ + Given store ID associated with this UCI, get volume ID as it is registered + on the cloud provider (e.g., 'vol-39890501') + """ + uci = model.UCI.get( self.uci_id ) + uci.refresh() + return uci.store[store_id].volume_id + + def get_all_stores( self ): + """ Returns all storage volumes' database objects associated with this UCI. """ + return model.CloudStore.filter( model.CloudStore.c.uci_id == self.uci_id ).all() +# svs = model.CloudStore.filter( model.CloudStore.c.uci_id == self.uci_id ).all() +# svl = [] # storage volume list +# for sv in svs: +# svl.append( sv.volume_id ) +# return svl + + def get_uci( self ): + """ Returns database object for given UCI. """ + uci = model.UCI.get( self.uci_id ) + uci.refresh() + return uci + + def uci_launch_time_set( self ): + uci = model.UCI.get( self.uci_id ) + uci.refresh() + return uci.launch_time + + def delete( self ): + uci = model.UCI.get( self.uci_id ) + uci.refresh() +# uci.delete() + uci.state = 'deleted' # for bookkeeping reasons, mark as deleted but don't actually delete. + uci.flush() + class JobWrapper( object ): """ Wraps a 'model.Job' with convience methods for running processes and @@ -675,6 +978,11 @@ else: log.error( "Unable to start unknown cloud provider: %s" %self.provider_name ) + def put( self, uci_wrapper ): + """ Put given request for UCI manipulation into provider's request queue.""" +# log.debug( "Adding UCI '%s' manipulation request into cloud manager's queue." % uci_wrapper.name ) + self.cloud_provider[self.provider_name].put( uci_wrapper ) + def createUCI( self, uci ): """ Createse User Configured Instance (UCI). Essentially, creates storage volume. @@ -682,13 +990,15 @@ log.debug( "Creating UCI '%s'" % uci.name ) self.cloud_provider[self.provider_name].createUCI( uci ) - def deleteUCI( self, uciName ): + def deleteUCI( self, uci ): """ Deletes UCI. NOTE that this implies deletion of any and all data associated with this UCI from the cloud. All data will be deleted. """ + log.debug( "Deleting UCI '%s'" % uci.name ) + self.cloud_provider[self.provider_name].deleteUCI( uci ) - def addStorageToUCI( self, uciName ): + def addStorageToUCI( self, uci ): """ Adds more storage to specified UCI """ def startUCI( self, uci ): @@ -706,17 +1016,16 @@ """ log.debug( "Stopping UCI '%s'" % uci.name ) self.cloud_provider[self.provider_name].stopUCI( uci ) + + def update( self ): + """ + Runs a global status update on all storage volumes and all instances whose UCI is + 'running' state. + Reason behind this method is to sync state of local DB and real world resources + """ +# log.debug( "Running global update" ) + self.cloud_provider[self.provider_name].update() - def put( self, job_wrapper ): - runner_name = ( job_wrapper.tool.job_runner.split(":", 1) )[0] - log.debug( "dispatching job %d to %s runner" %( job_wrapper.job_id, runner_name ) ) - self.cloud_provider[runner_name].put( job_wrapper ) - - def stop( self, job ): - runner_name = ( job.job_runner_name.split(":", 1) )[0] - log.debug( "stopping job %d in %s runner" %( job.id, runner_name ) ) - self.cloud_provider[runner_name].stop_job( job ) - def recover( self, job, job_wrapper ): runner_name = ( job.job_runner_name.split(":", 1) )[0] log.debug( "recovering job %d in %s runner" %( job.id, runner_name ) ) diff -r 7c438fd3cf4a -r 6aab50510e43 lib/galaxy/cloud/providers/eucalyptus.py --- a/lib/galaxy/cloud/providers/eucalyptus.py Fri Oct 02 18:31:32 2009 -0400 +++ b/lib/galaxy/cloud/providers/eucalyptus.py Wed Oct 14 19:20:11 2009 -0400 @@ -3,7 +3,10 @@ from datetime import datetime from galaxy import model # Database interaction class +from galaxy.model import mapping from galaxy.datatypes.data import nice_size +from Queue import Queue +from sqlalchemy import or_ import galaxy.eggs galaxy.eggs.require("boto") @@ -22,56 +25,86 @@ log.debug( "Using eucalyptus as default cloud provider." ) self.zone = "epc" self.key_pair = "galaxy-keypair" + self.queue = Queue() #TODO: Use multiple threads to process requests? + self.threads = [] + nworkers = 5 + log.info( "Starting eucalyptus cloud controller workers" ) + for i in range( nworkers ): + worker = threading.Thread( target=self.run_next ) + worker.start() + self.threads.append( worker ) + log.debug( "%d cloud workers ready", nworkers ) - - def get_connection( self, uci ): + def run_next( self ): + """Run the next job, waiting until one is available if necessary""" + cnt = 0 + while 1: + + uci_wrapper = self.queue.get() +# uci = uci_wrapper.get_uci() + log.debug( '[%d] uci name: %s' % ( cnt, uci_wrapper.get_name() ) ) + uci_state = uci_wrapper.get_state() + if uci_state is self.STOP_SIGNAL: + return + try: + if uci_state=="new": + log.debug( "Calling create UCI" ) + self.createUCI( uci_wrapper ) + elif uci_state=="deleting": + self.deleteUCI( uci_wrapper ) + elif uci_state=="submitted": + log.debug( "Calling start UCI" ) + self.startUCI( uci_wrapper ) + elif uci_state=="shutting-down": + self.stopUCI( uci_wrapper ) + except: + log.exception( "Uncaught exception executing request." ) + cnt += 1 + + def get_connection( self, uci_wrapper ): """ Establishes EC2 connection using user's default credentials """ log.debug( '##### Establishing cloud connection' ) -# creds = model.CloudUserCredentials.filter_by( user=user, defaultCred=True ).first() - a_key = uci.credentials.access_key - s_key = uci.credentials.secret_key # Amazon EC2 - #conn = EC2Connection( a_key, s_key ) + #conn = EC2Connection( uci_wrapper.get_access_key(), uci_wrapper.get_secret_key() ) + # Eucalyptus Public Cloud - # TODO: Add option in Galaxy config file to specify these values (i.e., for locally manages Eucalyptus deployments) + # TODO: Add option in Galaxy config file to specify these values (i.e., for locally managed Eucalyptus deployments) euca_region = RegionInfo( None, "eucalyptus", "mayhem9.cs.ucsb.edu" ) - conn = EC2Connection( aws_access_key_id=a_key, aws_secret_access_key=s_key, is_secure=False, port=8773, region=euca_region, path="/services/Eucalyptus" ) + conn = EC2Connection( aws_access_key_id=uci_wrapper.get_access_key(), aws_secret_access_key=uci_wrapper.get_secret_key(), is_secure=False, port=8773, region=euca_region, path="/services/Eucalyptus" ) return conn - def get_keypair_name( self, uci, conn ): + def set_keypair( self, uci_wrapper, conn ): """ Generate keypair using user's default credentials """ log.debug( "Getting user's keypair" ) kp = conn.get_key_pair( self.key_pair ) + instances = uci_wrapper.get_instances_indexes() try: - for i, inst in enumerate( uci.instance ): - uci.instance[i].keypair_name = kp.name + for inst in instances: + log.debug("inst: '%s'" % inst ) + uci_wrapper.set_key_pair( inst, kp.name ) return kp.name except AttributeError: # No keypair under this name exists so create it - log.debug( "No keypair found, creating keypair '%s'" % self.key_pair ) + log.info( "No keypair found, creating keypair '%s'" % self.key_pair ) kp = conn.create_key_pair( self.key_pair ) - for i, inst in enumerate( uci.instance ): - uci.instance[i].keypair_name = kp.name - uci.instance[i].keypair_material = kp.material - uci.flush() - # TODO: Store key_pair.material into instance table - this is the only time private key can be retrieved - # Actually, probably return key_pair to calling method and store name & key from there... - + for inst in instances: + uci_wrapper.set_key_pair( inst, kp.name, kp.material ) + return kp.name - def get_mi( self, type='small' ): + def get_mi_id( self, type ): """ Get appropriate machine image (mi) based on instance size. TODO: Dummy method - need to implement logic For valid sizes, see http://aws.amazon.com/ec2/instance-types/ """ - return model.CloudImage.filter( model.CloudImage.table.c.id==1 ).first() + return model.CloudImage.filter( model.CloudImage.table.c.id==1 ).first().image_id # def get_instances( self, uci ): # """ @@ -91,25 +124,31 @@ def shutdown( self ): """Attempts to gracefully shut down the monitor thread""" log.info( "sending stop signal to worker threads in eucalyptus cloud manager" ) - self.queue.put( self.STOP_SIGNAL ) + for i in range( len( self.threads ) ): + self.queue.put( self.STOP_SIGNAL ) log.info( "eucalyptus cloud manager stopped" ) + + def put( self, uci_wrapper ): + # Get rid of UCI from state description + state = uci_wrapper.get_state() + uci_wrapper.change_state( state.split('U')[0] ) # remove 'UCI' from end of state description (i.e., mark as accepted and ready for processing) + self.queue.put( uci_wrapper ) - def createUCI( self, uci ): + def createUCI( self, uci_wrapper ): """ Creates User Configured Instance (UCI). Essentially, creates storage volume on cloud provider and registers relevant information in Galaxy database. """ - conn = self.get_connection( uci ) + conn = self.get_connection( uci_wrapper ) # Temporary code - need to ensure user selects zone at UCI creation time! - if uci.store[0].availability_zone=='': + if uci_wrapper.get_store_availability_zone( 0 )=='': log.info( "Availability zone for storage volume was not selected, using default zone: %s" % self.zone ) - uci.store[0].availability_zone = self.zone - uci.store[0].flush() + uci_wrapper.set_store_availability_zone( 0, self.zone ) #TODO: check if volume associated with UCI already exists (if server crashed for example) and don't recreate it - log.debug( "Creating volume in zone '%s'..." % uci.store[0].availability_zone ) - vol = conn.create_volume( uci.store[0].size, uci.store[0].availability_zone, snapshot=None ) - uci.store[0].volume_id = vol.id + log.info( "Creating volume in zone '%s'..." % uci_wrapper.get_store_availability_zone( 0 ) ) + vol = conn.create_volume( uci_wrapper.get_store_size( 0 ), uci_wrapper.get_store_availability_zone( 0 ), snapshot=None ) + uci_wrapper.set_store_volume_id( 0, vol.id ) # Wait for a while to ensure volume was created # vol_status = vol.status @@ -125,111 +164,180 @@ # uci.flush() # return - uci.state = 'available' - uci.store[0].status = vol.status - uci.store[0].flush() - uci.flush() + # EPC does not allow creation of storage volumes (it deletes one as soon as it is created, so manually set uci_state here) + uci_wrapper.change_state( uci_state='available' ) + uci_wrapper.set_store_status( 0, vol.status ) - def deleteUCI( self, name ): + def deleteUCI( self, uci_wrapper ): """ Deletes UCI. NOTE that this implies deletion of any and all data associated with this UCI from the cloud. All data will be deleted. """ - + conn = self.get_connection( uci_wrapper ) + vl = [] # volume list + count = 0 # counter for checking if all volumes assoc. w/ UCI were deleted + + # Get all volumes assoc. w/ UCI, delete them from cloud as well as in local DB + vl = uci_wrapper.get_all_stores() + deletedList = [] + failedList = [] + for v in vl: + log.debug( "Deleting volume with id='%s'" % v.volume_id ) + if conn.delete_volume( v.volume_id ): + deletedList.append( v.volume_id ) + v.delete() + v.flush() + count += 1 + else: + failedList.append( v.volume_id ) + + # Delete UCI if all of associated + log.debug( "count=%s, len(vl)=%s" % (count, len( vl ) ) ) + if count == len( vl ): + uci_wrapper.delete() + else: + log.error( "Deleting following volume(s) failed: %s. However, these volumes were successfully deleted: %s. \ + MANUAL intervention and processing needed." % ( failedList, deletedList ) ) + uci_wrapper.change_state( uci_state="error" ) + def addStorageToUCI( self, name ): """ Adds more storage to specified UCI """ + + def dummyStartUCI( self, uci_wrapper ): - def startUCI( self, uci ): + uci = uci_wrapper.get_uci() + log.debug( "Would be starting instance '%s'" % uci.name ) + uci_wrapper.change_state( 'pending' ) +# log.debug( "Sleeping a bit... (%s)" % uci.name ) +# time.sleep(20) +# log.debug( "Woke up! (%s)" % uci.name ) + + def startUCI( self, uci_wrapper ): """ - Starts an instance of named UCI on the cloud. This implies, mounting of - storage and starting Galaxy instance. + Starts an instance of named UCI on the cloud. """ - conn = self.get_connection( uci ) + conn = self.get_connection( uci_wrapper ) +# + self.set_keypair( uci_wrapper, conn ) - uci.instance[0].keypair_name = self.get_keypair_name( uci, conn ) - mi = self.get_mi( uci.instance[0].type ) + i_indexes = uci_wrapper.get_instances_indexes() # Get indexes of i_indexes associated with this UCI -# log.debug( "mi: %s, mi.image_id: %s, uci.instance[0].keypair_name: %s" % ( mi, mi.image_id, uci.instance[0].keypair_name ) ) - uci.instance[0].image = mi + for i_index in i_indexes: + mi_id = self.get_mi_id( uci_wrapper.get_type( i_index ) ) + log.debug( "mi_id: %s, uci_wrapper.get_key_pair_name( i_index ): %s" % ( mi_id, uci_wrapper.get_key_pair_name( i_index ) ) ) + uci_wrapper.set_mi( i_index, mi_id ) + + # log.debug( '***** Setting up security group' ) + # If not existent, setup galaxy security group + # try: + # gSecurityGroup = conn.create_security_group('galaxy', 'Security group for Galaxy.') + # gSecurityGroup.authorize( 'tcp', 80, 80, '0.0.0.0/0' ) # Open HTTP port + # gSecurityGroup.authorize( 'tcp', 22, 22, '0.0.0.0/0' ) # Open SSH port + # except: + # pass + # sgs = conn.get_all_security_groups() + # for i in range( len( sgs ) ): + # if sgs[i].name == "galaxy": + # sg.append( sgs[i] ) + # break # only 1 security group w/ this name can exist, so continue + + log.debug( "***** Starting UCI instance '%s'" % uci_wrapper.get_name() ) + log.debug( 'Using following command: conn.run_instances( image_id=%s, key_name=%s )' % ( mi_id, uci_wrapper.get_key_pair_name( i_index ) ) ) + reservation = conn.run_instances( image_id=mi_id, key_name=uci_wrapper.get_key_pair_name( i_index ) ) + #reservation = conn.run_instances( image_id=instance.image, key_name=instance.keypair_name, security_groups=['galaxy'], instance_type=instance.type, placement=instance.availability_zone ) + l_time = datetime.utcnow() + uci_wrapper.set_launch_time( l_time, i_index=i_index ) # format_time( reservation.i_indexes[0].launch_time ) ) + if not uci_wrapper.uci_launch_time_set(): + uci_wrapper.set_uci_launch_time( l_time ) + uci_wrapper.set_reservation_id( i_index, str( reservation ).split(":")[1] ) + i_id = str( reservation.instances[0]).split(":")[1] + uci_wrapper.set_instance_id( i_index, i_id ) + s = reservation.instances[0].state # TODO: once more than a single instance will be started through single reservation, change this + uci_wrapper.change_state( s, i_id, s ) + log.debug( "UCI '%s' started, current state: %s" % ( uci_wrapper.get_name(), uci_wrapper.get_state() ) ) -# log.debug( '***** Setting up security group' ) - # If not existent, setup galaxy security group -# try: -# gSecurityGroup = conn.create_security_group('galaxy', 'Security group for Galaxy.') -# gSecurityGroup.authorize( 'tcp', 80, 80, '0.0.0.0/0' ) # Open HTTP port -# gSecurityGroup.authorize( 'tcp', 22, 22, '0.0.0.0/0' ) # Open SSH port -# except: -# pass -# sgs = conn.get_all_security_groups() -# for i in range( len( sgs ) ): -# if sgs[i].name == "galaxy": -# sg.append( sgs[i] ) -# break # only 1 security group w/ this name can exist, so continue - - log.debug( "***** Starting UCI instance '%s'" % uci.name ) -# log.debug( 'Using following command: conn.run_instances( image_id=%s, key_name=%s )' % ( uci.instance[0].image.image_id, uci.instance[0].keypair_name ) ) - reservation = conn.run_instances( image_id=uci.instance[0].image.image_id, key_name=uci.instance[0].keypair_name ) - #reservation = conn.run_instances( image_id=instance.image, key_name=instance.keypair_name, security_groups=['galaxy'], instance_type=instance.type, placement=instance.availability_zone ) - uci.instance[0].launch_time = datetime.utcnow() - uci.launch_time = uci.instance[0].launch_time - uci.instance[0].reservation_id = str( reservation ).split(":")[1] - uci.instance[0].instance_id = str( reservation.instances[0]).split(":")[1] - s = reservation.instances[0].state - uci.instance[0].state = s - uci.state = s - uci.instance[0].flush() - uci.flush() - # Wait until instance gets running and then update the DB - while s!="running": - log.debug( "Waiting on instance '%s' to start up (reservation ID: %s); current state: %s" % ( uci.instance[0].instance_id, uci.instance[0].reservation_id, s ) ) - time.sleep( 15 ) - s = reservation.instances[0].update() - - uci.instance[0].state = s - uci.state = s - uci.instance[0].public_dns = reservation.instances[0].dns_name - uci.instance[0].private_dns = reservation.instances[0].private_dns_name - uci.instance[0].flush() - uci.flush() +# # Wait until instance gets running and then update the DB +# while s!="running": +# log.debug( "Waiting on instance '%s' to start up (reservation ID: %s); current state: %s" % ( uci.instance[0].instance_id, uci.instance[0].reservation_id, s ) ) +# time.sleep( 15 ) +# s = reservation.i_indexes[0].update() +# +# # Update instance data in local DB +# uci.instance[0].state = s +# uci.instance[0].public_dns = reservation.i_indexes[0].dns_name +# uci.instance[0].private_dns = reservation.i_indexes[0].private_dns_name +# uci.instance[0].flush() +# # Update storage data in local DB w/ volume state info. NOTE that this only captures current volume state +# # and does not connect or wait on connection between instance and volume to be established +# vl = model.CloudStore.filter( model.CloudStore.c.uci_id == uci.id ).all() +# vols = [] +# for v in vl: +# vols.append( v.volume_id ) +# try: +# volumes = conn.get_all_volumes( vols ) +# for i, v in enumerate( volumes ): +# uci.store[i].i_id = v.instance_id +# uci.store[i].status = v.status +# uci.store[i].device = v.device +# uci.store[i].flush() +# except BotoServerError: +# log.debug( "Error getting volume(s) attached to instance. Volume status was not updated." ) +# +# uci.state = s +# uci.flush() - def stopUCI( self, uci ): + + def stopUCI( self, uci_wrapper): """ Stops all of cloud instances associated with named UCI. """ - conn = self.get_connection( uci ) - tl = [] # temination list + conn = self.get_connection( uci_wrapper ) - for i, inst in enumerate( uci.instance ): - tl.append( uci.instance[i].instance_id ) + # Get all instances associated with given UCI + il = uci_wrapper.get_instances_ids() # instance list +# log.debug( 'List of instances being terminated: %s' % il ) + rl = conn.get_all_instances( il ) # Reservation list associated with given instances - instList = conn.get_all_instances( tl ) -# log.debug( 'instList: %s' % instList ) - - for i, inst in enumerate( instList ): -# log.debug( 'inst: %s' % inst ) - log.debug( 'Before stop - inst.instances[0].update(): %s' % inst.instances[0].update() ) - inst.instances[0].stop() - log.debug( 'After stop - inst.instances[0].update(): %s' % inst.instances[0].update() ) - uci.instance[i].stop_time = datetime.utcnow() +# tState = conn.terminate_instances( il ) +# # TODO: Need to update instance stop time (for all individual instances) +# stop_time = datetime.utcnow() +# uci_wrapper.set_stop_time( stop_time ) - terminated=0 - while terminated!=len( instList ): - for i, inst in enumerate( instList ): - log.debug( "inst state: %s" % inst.instances[0].state ) - state = inst.instances[0].update() - if state=='terminated': - uci.instance[i].state = state - uci.instance[i].flush() - terminated += 1 - time.sleep ( 5 ) - - uci.state = 'available' - uci.launch_time = None - uci.flush() - - log.debug( "All instances for UCI '%s' were terminated." % uci.name ) + # Initiate shutdown of all instances under given UCI + cnt = 0 + stopped = [] + notStopped = [] + for r in rl: + for inst in r.instances: + log.debug( "Sending stop signal to instance '%s' associated with reservation '%s'." % ( inst, r ) ) + inst.stop() + uci_wrapper.set_stop_time( datetime.utcnow(), i_id=inst.id ) + uci_wrapper.change_state( instance_id=inst.id, i_state=inst.update() ) + stopped.append( inst ) + +# uci_wrapper.change_state( uci_state='available' ) + uci_wrapper.reset_uci_launch_time() + +# # Wait for all instances to actually terminate and update local DB +# terminated=0 +# while terminated!=len( rl ): +# for i, r in enumerate( rl ): +# log.debug( "r state: %s" % r.instances[0].state ) +# state = r.instances[0].update() +# if state=='terminated': +# uci.instance[i].state = state +# uci.instance[i].flush() +# terminated += 1 +# time.sleep ( 5 ) +# +# # Reset UCI state +# uci.state = 'available' +# uci.launch_time = None +# uci.flush() +# + log.debug( "All instances for UCI '%s' were terminated." % uci_wrapper.get_name() ) @@ -277,4 +385,144 @@ # session.flush() # trans.log_event( "User stopped cloud instance '%s'" % uci.name ) # trans.set_message( "Galaxy instance '%s' stopped." % uci.name ) - \ No newline at end of file + + def update( self ): + """ + Runs a global status update on all storage volumes and all instances whose UCI is in + 'running', 'pending', or 'shutting-down' state. + Reason behind this method is to sync state of local DB and real-world resources + """ + log.debug( "Running general status update for EPC UCIs." ) + instances = model.CloudInstance.filter( or_( model.CloudInstance.c.state=="running", model.CloudInstance.c.state=="pending", model.CloudInstance.c.state=="shutting-down" ) ).all() + for inst in instances: + log.debug( "Running general status update on instance '%s'" % inst.instance_id ) + self.updateInstance( inst ) + + stores = model.CloudStore.filter( or_( model.CloudStore.c.status=="in-use", model.CloudStore.c.status=="creating" ) ).all() + for store in stores: + log.debug( "Running general status update on store '%s'" % store.volume_id ) + self.updateStore( store ) + + def updateInstance( self, inst ): + + # Get credentials associated wit this instance + uci_id = inst.uci_id + uci = model.UCI.get( uci_id ) + uci.refresh() + a_key = uci.credentials.access_key + s_key = uci.credentials.secret_key + # Get connection + euca_region = RegionInfo( None, "eucalyptus", "mayhem9.cs.ucsb.edu" ) + conn = EC2Connection( aws_access_key_id=a_key, aws_secret_access_key=s_key, is_secure=False, port=8773, region=euca_region, path="/services/Eucalyptus" ) + # Get reservations handle for given instance + rl= conn.get_all_instances( [inst.instance_id] ) + # Because EPC deletes references to reservations after a short while after instances have terminated, getting an empty list as a response to a query + # typically means the instance has successfully shut down but the check was not performed in short enough amount of time. As a result, below code simply + # marks given instance as having terminated. Note that an instance might have also crashed and this code will not catch the difference... + if len( rl ) == 0: + log.info( "Instance ID '%s' was not found by the cloud provider. Instance might have crashed or otherwise been terminated." % inst.instance_id ) + inst.state = 'terminated' + uci.state = 'available' + uci.launch_time = None + inst.flush() + uci.flush() + # Update instance status in local DB with info from cloud provider + for r in rl: + for i, cInst in enumerate( r.instances ): + s = cInst.update() + log.debug( "Checking state of cloud instance '%s' associated with reservation '%s'. State='%s'" % ( cInst, r, s ) ) + if s != inst.state: + inst.state = s + inst.flush() + if s == 'terminated': # After instance has shut down, ensure UCI is marked as 'available' + uci.state = 'available' + uci.flush() + if s != uci.state and s != 'terminated': + # Making sure state of UCI is updated. Once multiple instances become associated with single UCI, this will need to be changed. + uci.state = s + uci.flush() + if cInst.public_dns_name != inst.public_dns: + inst.public_dns = cInst.public_dns_name + inst.flush() + if cInst.private_dns_name != inst.private_dns: + inst.private_dns = cInst.private_dns_name + inst.flush() + + def updateStore( self, store ): + # Get credentials associated wit this store + uci_id = store.uci_id + uci = model.UCI.get( uci_id ) + uci.refresh() + a_key = uci.credentials.access_key + s_key = uci.credentials.secret_key + # Get connection + euca_region = RegionInfo( None, "eucalyptus", "mayhem9.cs.ucsb.edu" ) + conn = EC2Connection( aws_access_key_id=a_key, aws_secret_access_key=s_key, is_secure=False, port=8773, region=euca_region, path="/services/Eucalyptus" ) + # Get reservations handle for given store + vl = conn.get_all_volumes( [store.volume_id] ) +# log.debug( "Store '%s' vl: '%s'" % ( store.volume_id, vl ) ) + # Update store status in local DB with info from cloud provider + if store.status != vl[0].status: + store.status = vl[0].status + store.flush() + if store.i_id != vl[0].instance_id: + store.i_id = vl[0].instance_id + store.flush() + if store.attach_time != vl[0].attach_time: + store.attach_time = vl[0].attach_time + store.flush() + if store.device != vl[0].device: + store.device = vl[0].device + store.flush() + + def updateUCI( self, uci ): + """ + Runs a global status update on all storage volumes and all instances that are + associated with specified UCI + """ + conn = self.get_connection( uci ) + + # Update status of storage volumes + vl = model.CloudStore.filter( model.CloudInstance.c.uci_id == uci.id ).all() + vols = [] + for v in vl: + vols.append( v.volume_id ) + try: + volumes = conn.get_all_volumes( vols ) + for i, v in enumerate( volumes ): + uci.store[i].i_id = v.instance_id + uci.store[i].status = v.status + uci.store[i].device = v.device + uci.store[i].flush() + except: + log.debug( "Error updating status of volume(s) associated with UCI '%s'. Status was not updated." % uci.name ) + pass + + # Update status of instances + il = model.CloudInstance.filter_by( uci_id=uci.id ).filter( model.CloudInstance.c.state != 'terminated' ).all() + instanceList = [] + for i in il: + instanceList.append( i.instance_id ) + log.debug( 'instanceList: %s' % instanceList ) + try: + reservations = conn.get_all_instances( instanceList ) + for i, r in enumerate( reservations ): + uci.instance[i].state = r.instances[0].update() + log.debug('updating instance %s; status: %s' % ( uci.instance[i].instance_id, uci.instance[i].state ) ) + uci.state = uci.instance[i].state + uci.instance[i].public_dns = r.instances[0].dns_name + uci.instance[i].private_dns = r.instances[0].private_dns_name + uci.instance[i].flush() + uci.flush() + except: + log.debug( "Error updating status of instances associated with UCI '%s'. Instance status was not updated." % uci.name ) + pass + + # --------- Helper methods ------------ + + def format_time( time ): + dict = {'T':' ', 'Z':''} + for i, j in dict.iteritems(): + time = time.replace(i, j) + return time + \ No newline at end of file diff -r 7c438fd3cf4a -r 6aab50510e43 lib/galaxy/model/mapping.py --- a/lib/galaxy/model/mapping.py Fri Oct 02 18:31:32 2009 -0400 +++ b/lib/galaxy/model/mapping.py Wed Oct 14 19:20:11 2009 -0400 @@ -965,7 +965,7 @@ properties=dict( user=relation( User ), credentials=relation( CloudUserCredentials ), instance=relation( CloudInstance, backref='uci' ), - store=relation( CloudStore, backref='uci' ) + store=relation( CloudStore, backref='uci', cascade='all, delete-orphan' ) ) ) assign_mapper( context, CloudInstance, CloudInstance.table, diff -r 7c438fd3cf4a -r 6aab50510e43 lib/galaxy/model/migrate/versions/0014_cloud_tables.py --- a/lib/galaxy/model/migrate/versions/0014_cloud_tables.py Fri Oct 02 18:31:32 2009 -0400 +++ b/lib/galaxy/model/migrate/versions/0014_cloud_tables.py Wed Oct 14 19:20:11 2009 -0400 @@ -82,7 +82,10 @@ CloudImage_table.create() except Exception, e: log.debug( "Creating cloud_image table failed. Table probably exists already." ) - UCI_table.create() + try: + UCI_table.create() + except Exception, e: + log.debug( "Creating UCI table failed. Table probably exists already." ) CloudInstance_table.create() CloudStore_table.create() try: @@ -115,7 +118,8 @@ log.debug( "Dropping cloud_user_credentials table failed: %s" % str( e ) ) try: - UCI_table.drop() + log.debug( "Would drop UCI table." ) +# UCI_table.drop() except Exception, e: log.debug( "Dropping UCI table failed: %s" % str( e ) ) diff -r 7c438fd3cf4a -r 6aab50510e43 lib/galaxy/web/controllers/cloud.py --- a/lib/galaxy/web/controllers/cloud.py Fri Oct 02 18:31:32 2009 -0400 +++ b/lib/galaxy/web/controllers/cloud.py Wed Oct 14 19:20:11 2009 -0400 @@ -56,25 +56,39 @@ liveInstances = trans.sa_session.query( model.UCI ) \ .filter_by( user=user ) \ - .filter( or_( model.UCI.c.state=="running", model.UCI.c.state=="pending", model.UCI.c.state=="terminating" ) ) \ - .order_by( desc( model.UCI.c.launch_time ) ) \ + .filter( or_( model.UCI.c.state=="running", + model.UCI.c.state=="pending", + model.UCI.c.state=="submitted", + model.UCI.c.state=="submittedUCI", + model.UCI.c.state=="shutting-down", + model.UCI.c.state=="shutting-downUCI" ) ) \ + .order_by( desc( model.UCI.c.update_time ) ) \ .all() prevInstances = trans.sa_session.query( model.UCI ) \ .filter_by( user=user ) \ - .filter( or_( model.UCI.c.state=="available", model.UCI.c.state=="new", model.UCI.c.state=="error", model.UCI.c.state=="submitted" ) ) \ + .filter( or_( model.UCI.c.state=="available", + model.UCI.c.state=="new", + model.UCI.c.state=="newUCI", + model.UCI.c.state=="error", + model.UCI.c.state=="deleting", + model.UCI.c.state=="deletingUCI" ) ) \ .order_by( desc( model.UCI.c.update_time ) ) \ .all() # Check after update there are instances in pending state; if so, display message # TODO: Auto-refresh once instance is running pendingInstances = trans.sa_session.query( model.UCI ) \ - .filter_by( user=user, state="pending" ) \ + .filter_by( user=user ) \ + .filter( or_( model.UCI.c.state=="pending" , \ + model.UCI.c.state=="submitted" , \ + model.UCI.c.state=="submittedUCI" ) ) \ .all() if pendingInstances: trans.set_message( "Galaxy instance started. NOTE: Please wait about 3-5 minutes for the instance to " "start up and then refresh this page. A button to connect to the instance will then appear alongside " "instance description." ) + return trans.fill_template( "cloud/configure_cloud.mako", cloudCredentials = cloudCredentials, liveInstances = liveInstances, @@ -113,7 +127,13 @@ uci = get_uci( trans, id ) stores = get_stores( trans, uci ) # log.debug(self.app.config.job_working_directory) - if len(stores) is not 0: + if ( len(stores) is not 0 ) and \ + ( uci.state != 'submitted' ) and \ + ( uci.state != 'submittedUCI' ) and \ + ( uci.state != 'pending' ) and \ + ( uci.state != 'deleting' ) and \ + ( uci.state != 'deletingUCI' ) and \ + ( uci.state != 'error' ): instance = model.CloudInstance() instance.user = user instance.image = mi @@ -146,7 +166,7 @@ # instance.instance_id = str( reservation.instances[0]).split(":")[1] # instance.state = "pending" # instance.state = reservation.instances[0].state - uci.state = 'submitted' + uci.state = 'submittedUCI' # Persist session = trans.sa_session @@ -155,8 +175,14 @@ session.flush() trans.log_event ("User initiated starting of cloud instance '%s'." % uci.name ) + trans.set_message( "Galaxy instance started. NOTE: Please wait about 3-5 minutes for the instance to " + "start up and then refresh this page. A button to connect to the instance will then appear alongside " + "instance description." ) return self.list( trans ) + trans.show_error_message( "Cannot start instance that is in state '%s'." % uci.state ) + return self.list( trans ) + # return trans.show_form( # web.FormBuilder( web.url_for(), "Start instance size", submit_text="Start" ) # .add_input( "radio","Small","size", value='small' ) @@ -170,7 +196,7 @@ Stop a cloud UCI instance. This implies stopping Galaxy server and disconnecting/unmounting relevant file system(s). """ uci = get_uci( trans, id ) - uci.state = 'terminating' + uci.state = 'shutting-downUCI' session = trans.sa_session # session.save_or_update( stores ) session.save_or_update( uci ) @@ -234,32 +260,38 @@ any and all storage associated with this UCI! """ uci = get_uci( trans, id ) - dbInstances = get_instances( trans, uci ) #TODO: handle list! - conn = get_connection( trans ) - session = trans.sa_session + if ( uci.state != 'deletingUCI' ) and ( uci.state != 'deleting' ) and ( uci.state != 'error' ): + name = uci.name + uci.state = "deletingUCI" + # dbInstances = get_instances( trans, uci ) #TODO: handle list! + # + # conn = get_connection( trans ) + session = trans.sa_session + # + # # Delete volume(s) associated with given uci + # stores = get_stores( trans, uci ) + # for i, store in enumerate( stores ): + # log.debug( "Deleting volume '%s' that is associated with UCI '%s'." % ( store.volume_id, uci.name ) ) + # volStat = None + # try: + # volStat = conn.delete_volume( store.volume_id ) + # except: + # log.debug ( 'Error deleting volume %s' % store.volume_id ) + # + # if volStat: + # session.delete( store ) + # + # # Delete UCI from table + # uciName = uci.name # Store name for logging + # session.delete( uci ) + + session.flush() + trans.log_event( "User deleted cloud instance '%s'" % name ) + trans.set_message( "Galaxy instance '%s' marked for deletion." % name ) + return self.list( trans ) - # Delete volume(s) associated with given uci - stores = get_stores( trans, uci ) - for i, store in enumerate( stores ): - log.debug( "Deleting volume '%s' that is associated with UCI '%s'." % ( store.volume_id, uci.name ) ) - volStat = None - try: - volStat = conn.delete_volume( store.volume_id ) - except: - log.debug ( 'Error deleting volume %s' % store.volume_id ) - - if volStat: - session.delete( store ) - - # Delete UCI from table - uciName = uci.name # Store name for logging - session.delete( uci ) - - session.flush() - trans.log_event( "User deleted cloud instance '%s'" % uciName ) - trans.set_message( "Galaxy instance '%s' deleted." % uciName ) - + trans.set_message( "Instance '%s' is already marked for deletion." % uci.name ) return self.list( trans ) @web.expose @@ -280,7 +312,7 @@ """ inst_error = vol_error = cred_error = None user = trans.get_user() - # TODO: Hack until present user w/ bullet list w/ registered credentials + # TODO: Hack until present user w/ bullet list w/ registered credentials storedCreds = trans.sa_session.query( model.CloudUserCredentials ) \ .filter_by( user=user ).all() credsMatch = False @@ -293,8 +325,7 @@ try: if len( instanceName ) > 255: inst_error = "Instance name exceeds maximum allowable length." - elif trans.app.model.UCI.filter( - trans.app.model.UCI.table.c.name==instanceName ).first(): + elif trans.app.model.UCI.filter( and_( trans.app.model.UCI.table.c.name==instanceName, trans.app.model.UCI.table.c.state!='deleted' ) ).first(): inst_error = "An instance with that name already exist." elif int( volSize ) > 1000: vol_error = "Volume size cannot exceed 1000GB. You must specify an integer between 1 and 1000." @@ -310,7 +341,7 @@ trans.app.model.CloudUserCredentials.table.c.name==credName ).first() uci.user= user uci.total_size = volSize # This is OK now because new instance is being created. - uci.state = "new" # Valid states include: "new, "available", "running", "pending", "submitted", "terminating", or "error" + uci.state = "newUCI" storage = model.CloudStore() storage.user = user storage.uci = uci @@ -403,15 +434,20 @@ Add user's cloud credentials stored under name `credName`. """ user = trans.get_user() - cred_error = accessKey_error = secretKey_error = provider_error = None + error = {} + if credName: if len( credName ) > 255: - cred_error = "Credentials name exceeds maximum allowable length." + error['cred_error'] = "Credentials name exceeds maximum allowable length." elif trans.app.model.CloudUserCredentials.filter( trans.app.model.CloudUserCredentials.table.c.name==credName ).first(): - cred_error = "Credentials with that name already exist." + error['cred_error'] = "Credentials with that name already exist." elif ( ( providerName.lower()!='ec2' ) and ( providerName.lower()!='eucalyptus' ) ): - provider_error = "You specified an unsupported cloud provider." + error['provider_error'] = "You specified an unsupported cloud provider." + elif accessKey=='' or len( accessKey ) > 255: + error['access_key_error'] = "Access key much be between 1 and 255 characters long." + elif secretKey=='' or len( secretKey ) > 255: + error['secret_key_error'] = "Secret key much be between 1 and 255 characters long." else: # Create new user stored credentials credentials = model.CloudUserCredentials() @@ -430,12 +466,21 @@ # if defaultCred: # self.makeDefault( trans, credentials.id) return self.list( trans ) - return trans.show_form( - web.FormBuilder( web.url_for(), "Add credentials", submit_text="Add" ) - .add_text( "credName", "Credentials name", value="Unnamed credentials", error=cred_error ) - .add_text( "providerName", "Cloud provider name", value="ec2 or eucalyptus", error=provider_error ) - .add_text( "accessKey", "Access key", value='', error=accessKey_error ) - .add_password( "secretKey", "Secret key", value='', error=secretKey_error ) ) + + return trans.fill_template( "cloud/add_credentials.mako", \ + credName = credName, \ + providerName = providerName, \ + accessKey = accessKey, \ + secretKey = secretKey, \ + error = error + ) + +# return trans.show_form( +# web.FormBuilder( web.url_for(), "Add credentials", submit_text="Add" ) +# .add_text( "credName", "Credentials name", value="Unnamed credentials", error=cred_error ) +# .add_text( "providerName", "Cloud provider name", value="ec2 or eucalyptus", error=provider_error ) +# .add_text( "accessKey", "Access key", value='', error=accessKey_error ) +# .add_password( "secretKey", "Secret key", value='', error=secretKey_error ) ) @web.expose @web.require_login( "view credentials" ) @@ -1018,7 +1063,7 @@ # Retrieve cloud instance based on passed instance id. get_all_instances( idLst ) method returns reservation ID. Because # we are passing only 1 ID, we can retrieve only the first element of the returning list. Furthermore, because (for now!) # only 1 instance corresponds each individual reservation, grab only the first element of the returned list of instances. - cloudInstance = conn.get_all_instances( idLst )[0].instances[0] + cloudInstance = conn.get_all_instances( [instance_id] )[0].instances[0] return cloudInstance def get_connection( trans, credName ): diff -r 7c438fd3cf4a -r 6aab50510e43 templates/cloud/add_credentials.mako --- /dev/null Thu Jan 01 00:00:00 1970 +0000 +++ b/templates/cloud/add_credentials.mako Wed Oct 14 19:20:11 2009 -0400 @@ -0,0 +1,97 @@ +<% _=n_ %> +<%inherit file="/base.mako"/> +<%def name="title()">Add credentials</%def> + +<%def name="javascripts()"> +${parent.javascripts()} +<script type="text/javascript"> +$(function(){ + $("input:text:first").focus(); +}) +</script> +</%def> + +%if header: + ${header} +%endif + +<div class="form"> + <div class="form-title">Add credentials</div> + <div class="form-body"> + <form name="Add credentials" action="${h.url_for( action='add' )}" method="post" > + + <% + cls = "form-row" + if error.has_key('cred_error'): + cls += " form-row-error" + %> + <div class="${cls}"> + <label>Credentials name:</label> + <div class="form-row-input"> + <input type="text" name="credName" value="Unnamed credentials" size="40"> + </div> + %if error.has_key('cred_error'): + <div class="form-row-error-message">${error['cred_error']}</div> + %endif + <div style="clear: both"></div> + </div> + + + <% + cls = "form-row" + if error.has_key('provider_error'): + cls += " form-row-error" + %> + <div class="${cls}"> + <label>Cloud provider name:</label> + <div class="form-row-input"> + <select name="providerName"> + <option value="ec2">Amazon EC2</option> + <option value="eucalyptus">Eucalpytus Public Cloud (EPC)</option> + </select> + </div> + %if error.has_key('provider_error'): + <div class="form-row-error-message">${error['provider_error']}</div> + %endif + <div style="clear: both"></div> + </div> + + <% + cls = "form-row" + if error.has_key('access_key_error'): + cls += " form-row-error" + %> + <div class="${cls}"> + <label>Access key:</label> + <div class="form-row-input"> + <input type="text" name="accessKey" value="" size="40"> + </div> + %if error.has_key('access_key_error'): + <div class="form-row-error-message">${error['access_key_error']}</div> + %endif + <div style="clear: both"></div> + </div> + + + <% + cls = "form-row" + if error.has_key('secret_key_error'): + cls += " form-row-error" + %> + <div class="${cls}"> + <label>Secret key:</label> + <div class="form-row-input"> + <input type="password" name="secretKey" value="" size="40"> + </div> + %if error.has_key('secret_key_error'): + <div class="form-row-error-message">${error['secret_key_error']}</div> + %endif + <div style="clear: both"></div> + </div> + + + <div class="form-row"><input type="submit" value="Add"></div> + + </form> + </div> +</div>

1 0

[hg] galaxy 3067: Added most of the functionality for interactio...
by Greg Von Kuster 23 Nov '09

23 Nov '09

details: http://www.bx.psu.edu/hg/galaxy/rev/7c438fd3cf4a changeset: 3067:7c438fd3cf4a user: Enis Afgan <afgane(a)gmail.com> date: Fri Oct 02 18:31:32 2009 -0400 description: Added most of the functionality for interaction with Eucalyptus cloud diffstat: lib/galaxy/app.py | 4 +- lib/galaxy/cloud/__init__.py | 262 ++++++++++++--------- lib/galaxy/cloud/providers/eucalyptus.py | 275 +++++++++++++++++++--- lib/galaxy/model/mapping.py | 4 +- lib/galaxy/model/migrate/versions/0014_cloud_tables.py | 16 +- lib/galaxy/web/controllers/cloud.py | 160 ++++++------ 6 files changed, 465 insertions(+), 256 deletions(-) diffs (986 lines): diff -r c9c9adf06e9d -r 7c438fd3cf4a lib/galaxy/app.py --- a/lib/galaxy/app.py Wed Sep 30 17:57:11 2009 -0400 +++ b/lib/galaxy/app.py Fri Oct 02 18:31:32 2009 -0400 @@ -1,6 +1,6 @@ import sys, os, atexit -from galaxy import config, jobs, util, tools, web +from galaxy import config, jobs, util, tools, web, cloud from galaxy.tracks import store from galaxy.web import security import galaxy.model @@ -68,6 +68,8 @@ # FIXME: These are exposed directly for backward compatibility self.job_queue = self.job_manager.job_queue self.job_stop_queue = self.job_manager.job_stop_queue + # Start the cloud manager + self.cloud_manager = cloud.CloudManager( self ) # Track Store self.track_store = store.TrackStoreManager( self.config.track_store_path ) diff -r c9c9adf06e9d -r 7c438fd3cf4a lib/galaxy/cloud/__init__.py --- a/lib/galaxy/cloud/__init__.py Wed Sep 30 17:57:11 2009 -0400 +++ b/lib/galaxy/cloud/__init__.py Fri Oct 02 18:31:32 2009 -0400 @@ -6,7 +6,7 @@ from galaxy.datatypes.tabular import * from galaxy.datatypes.interval import * from galaxy.datatypes import metadata -#from util import Bunch +from galaxy.util.bunch import Bunch import pkg_resources pkg_resources.require( "PasteDeploy" ) @@ -34,7 +34,7 @@ # The dispatcher manager underlying cloud instances self.provider = DefaultCloudProvider( app ) # Monitor for updating status of cloud instances -# self.cloud_monitor = CloudMonitor( self.config, self.provider ) + self.cloud_monitor = CloudMonitor( self.app, self.provider ) # self.job_stop_queue = JobStopQueue( app, self.dispatcher ) else: self.job_queue = self.job_stop_queue = NoopCloudMonitor() @@ -43,32 +43,32 @@ self.cloud_monitor.shutdown() # self.job_stop_queue.shutdown() - def createUCI( self, user, name, storage_size, zone=None): - """ - Createse User Configured Instance (UCI). Essentially, creates storage volume. - """ - self.provider.createUCI( user, name, storage_size, zone ) - - def deleteUCI( self, name ): - """ - Deletes UCI. NOTE that this implies deletion of any and all data associated - with this UCI from the cloud. All data will be deleted. - """ - - def addStorageToUCI( self, name ): - """ Adds more storage to specified UCI """ - - def startUCI( self, name, type ): - """ - Starts an instance of named UCI on the cloud. This implies, mounting of - storage and starting Galaxy instance. - """ - - def stopUCI( self, name ): - """ - Stops cloud instance associated with named UCI. This also implies - stopping of Galaxy and unmounting of the file system. - """ +# def createUCI( self, user, name, storage_size, zone=None): +# """ +# Createse User Configured Instance (UCI). Essentially, creates storage volume. +# """ +# self.provider.createUCI( user, name, storage_size, zone ) +# +# def deleteUCI( self, name ): +# """ +# Deletes UCI. NOTE that this implies deletion of any and all data associated +# with this UCI from the cloud. All data will be deleted. +# """ +# +# def addStorageToUCI( self, name ): +# """ Adds more storage to specified UCI """ +# +# def startUCI( self, name, type ): +# """ +# Starts an instance of named UCI on the cloud. This implies, mounting of +# storage and starting Galaxy instance. +# """ +# +# def stopUCI( self, name ): +# """ +# Stops cloud instance associated with named UCI. This also implies +# stopping of Galaxy and unmounting of the file system. +# """ class Sleeper( object ): """ @@ -149,8 +149,8 @@ while self.running: try: - #self.__monitor_step() - log.debug ( "would be calling monitor_step" ) +# log.debug( "Calling monitor_step" ) + self.__monitor_step() except: log.exception( "Exception in cloud manager monitor_step" ) # Sleep @@ -171,88 +171,110 @@ session = mapping.Session() # Pull all new jobs from the queue at once new_jobs = [] - if self.track_jobs_in_database: - for j in session.query( model.Job ).options( lazyload( "external_output_metadata" ), lazyload( "parameters" ) ).filter( model.Job.c.state == model.Job.states.NEW ).all(): - job = JobWrapper( j, self.app.toolbox.tools_by_id[ j.tool_id ], self ) - new_jobs.append( job ) - else: - try: - while 1: - message = self.queue.get_nowait() - if message is self.STOP_SIGNAL: - return - # Unpack the message - job_id, tool_id = message - # Create a job wrapper from it - job_entity = session.query( model.Job ).get( job_id ) - job = JobWrapper( job_entity, self.app.toolbox.tools_by_id[ tool_id ], self ) - # Append to watch queue - new_jobs.append( job ) - except Empty: - pass - # Iterate over new and waiting jobs and look for any that are - # ready to run - new_waiting = [] - for job in ( new_jobs + self.waiting ): - try: - # Clear the session for each job so we get fresh states for - # job and all datasets - session.clear() - # Get the real job entity corresponding to the wrapper (if we - # are tracking in the database this is probably cached in - # the session from the origianl query above) - job_entity = session.query( model.Job ).get( job.job_id ) - # Check the job's dependencies, requeue if they're not done - job_state = self.__check_if_ready_to_run( job, job_entity ) - if job_state == JOB_WAIT: - if not self.track_jobs_in_database: - new_waiting.append( job ) - elif job_state == JOB_ERROR: - log.info( "job %d ended with an error" % job.job_id ) - elif job_state == JOB_INPUT_ERROR: - log.info( "job %d unable to run: one or more inputs in error state" % job.job_id ) - elif job_state == JOB_INPUT_DELETED: - log.info( "job %d unable to run: one or more inputs deleted" % job.job_id ) - elif job_state == JOB_READY: - # If special queuing is enabled, put the ready jobs in the special queue - if self.use_policy : - self.squeue.put( job ) - log.debug( "job %d put in policy queue" % job.job_id ) - else: # or dispatch the job directly - self.dispatcher.put( job ) - log.debug( "job %d dispatched" % job.job_id) - elif job_state == JOB_DELETED: - msg = "job %d deleted by user while still queued" % job.job_id - job.info = msg - log.debug( msg ) - elif job_state == JOB_ADMIN_DELETED: - job.fail( job_entity.info ) - log.info( "job %d deleted by admin while still queued" % job.job_id ) - else: - msg = "unknown job state '%s' for job %d" % ( job_state, job.job_id ) - job.info = msg - log.error( msg ) - except Exception, e: - job.info = "failure running job %d: %s" % ( job.job_id, str( e ) ) - log.exception( "failure running job %d" % job.job_id ) - # Update the waiting list - self.waiting = new_waiting - # If special (e.g. fair) scheduling is enabled, dispatch all jobs - # currently in the special queue - if self.use_policy : - while 1: - try: - sjob = self.squeue.get() - self.dispatcher.put( sjob ) - log.debug( "job %d dispatched" % sjob.job_id ) - except Empty: - # squeue is empty, so stop dispatching - break - except Exception, e: # if something else breaks while dispatching - job.fail( "failure running job %d: %s" % ( sjob.job_id, str( e ) ) ) - log.exception( "failure running job %d" % sjob.job_id ) - # Done with the session - mapping.Session.remove() + new_instances = [] + new_UCIs = [] + stop_UCIs = [] + +# for r in session.query( model.cloud_instance ).filter( model.cloud_instance.s.state == model.cloud_instance.states.NEW ).all(): +# new_instances + + for r in session.query( model.UCI ).filter( model.UCI.c.state == "new" ).all(): + new_UCIs.append( r ) + for r in new_UCIs: + self.provider.createUCI( r ) + + for r in session.query( model.UCI ).filter( model.UCI.c.state == "submitted" ).all(): + new_instances.append( r ) + for r in new_instances: + self.provider.startUCI( r ) + + for r in session.query( model.UCI ).filter( model.UCI.c.state == "terminating" ).all(): + stop_UCIs.append( r ) + for r in stop_UCIs: + self.provider.stopUCI( r ) + +# if self.track_jobs_in_database: +# for j in session.query( model.Job ).options( lazyload( "external_output_metadata" ), lazyload( "parameters" ) ).filter( model.Job.c.state == model.Job.states.NEW ).all(): +# job = JobWrapper( j, self.app.toolbox.tools_by_id[ j.tool_id ], self ) +# new_jobs.append( job ) +# else: +# try: +# while 1: +# message = self.queue.get_nowait() +# if message is self.STOP_SIGNAL: +# return +# # Unpack the message +# job_id, tool_id = message +# # Create a job wrapper from it +# job_entity = session.query( model.Job ).get( job_id ) +# job = JobWrapper( job_entity, self.app.toolbox.tools_by_id[ tool_id ], self ) +# # Append to watch queue +# new_jobs.append( job ) +# except Empty: +# pass +# # Iterate over new and waiting jobs and look for any that are +# # ready to run +# new_waiting = [] +# for job in ( new_jobs + self.waiting ): +# try: +# # Clear the session for each job so we get fresh states for +# # job and all datasets +# session.clear() +# # Get the real job entity corresponding to the wrapper (if we +# # are tracking in the database this is probably cached in +# # the session from the origianl query above) +# job_entity = session.query( model.Job ).get( job.job_id ) +# # Check the job's dependencies, requeue if they're not done +# job_state = self.__check_if_ready_to_run( job, job_entity ) +# if job_state == JOB_WAIT: +# if not self.track_jobs_in_database: +# new_waiting.append( job ) +# elif job_state == JOB_ERROR: +# log.info( "job %d ended with an error" % job.job_id ) +# elif job_state == JOB_INPUT_ERROR: +# log.info( "job %d unable to run: one or more inputs in error state" % job.job_id ) +# elif job_state == JOB_INPUT_DELETED: +# log.info( "job %d unable to run: one or more inputs deleted" % job.job_id ) +# elif job_state == JOB_READY: +# # If special queuing is enabled, put the ready jobs in the special queue +# if self.use_policy : +# self.squeue.put( job ) +# log.debug( "job %d put in policy queue" % job.job_id ) +# else: # or dispatch the job directly +# self.dispatcher.put( job ) +# log.debug( "job %d dispatched" % job.job_id) +# elif job_state == JOB_DELETED: +# msg = "job %d deleted by user while still queued" % job.job_id +# job.info = msg +# log.debug( msg ) +# elif job_state == JOB_ADMIN_DELETED: +# job.fail( job_entity.info ) +# log.info( "job %d deleted by admin while still queued" % job.job_id ) +# else: +# msg = "unknown job state '%s' for job %d" % ( job_state, job.job_id ) +# job.info = msg +# log.error( msg ) +# except Exception, e: +# job.info = "failure running job %d: %s" % ( job.job_id, str( e ) ) +# log.exception( "failure running job %d" % job.job_id ) +# # Update the waiting list +# self.waiting = new_waiting +# # If special (e.g. fair) scheduling is enabled, dispatch all jobs +# # currently in the special queue +# if self.use_policy : +# while 1: +# try: +# sjob = self.squeue.get() +# self.dispatcher.put( sjob ) +# log.debug( "job %d dispatched" % sjob.job_id ) +# except Empty: +# # squeue is empty, so stop dispatching +# break +# except Exception, e: # if something else breaks while dispatching +# job.fail( "failure running job %d: %s" % ( sjob.job_id, str( e ) ) ) +# log.exception( "failure running job %d" % sjob.job_id ) +# # Done with the session +# mapping.Session.remove() def __check_if_ready_to_run( self, job_wrapper, job ): """ @@ -653,12 +675,12 @@ else: log.error( "Unable to start unknown cloud provider: %s" %self.provider_name ) - def createUCI( self, user, uciName, storage_size, zone=None): + def createUCI( self, uci ): """ Createse User Configured Instance (UCI). Essentially, creates storage volume. """ - log.debug( "Creating UCI %s" % uciName ) - self.cloud_provider[self.provider_name].createUCI( user, uciName, storage_size, zone ) + log.debug( "Creating UCI '%s'" % uci.name ) + self.cloud_provider[self.provider_name].createUCI( uci ) def deleteUCI( self, uciName ): """ @@ -669,18 +691,22 @@ def addStorageToUCI( self, uciName ): """ Adds more storage to specified UCI """ - def startUCI( self, uciName, type ): + def startUCI( self, uci ): """ Starts an instance of named UCI on the cloud. This implies, mounting of storage and starting Galaxy instance. """ + log.debug( "Starting UCI '%s'" % uci.name ) + self.cloud_provider[self.provider_name].startUCI( uci ) - def stopUCI( self, uciName ): + def stopUCI( self, uci ): """ Stops cloud instance associated with named UCI. This also implies stopping of Galaxy and unmounting of the file system. """ - + log.debug( "Stopping UCI '%s'" % uci.name ) + self.cloud_provider[self.provider_name].stopUCI( uci ) + def put( self, job_wrapper ): runner_name = ( job_wrapper.tool.job_runner.split(":", 1) )[0] log.debug( "dispatching job %d to %s runner" %( job_wrapper.job_id, runner_name ) ) diff -r c9c9adf06e9d -r 7c438fd3cf4a lib/galaxy/cloud/providers/eucalyptus.py --- a/lib/galaxy/cloud/providers/eucalyptus.py Wed Sep 30 17:57:11 2009 -0400 +++ b/lib/galaxy/cloud/providers/eucalyptus.py Fri Oct 02 18:31:32 2009 -0400 @@ -1,12 +1,10 @@ -import subprocess, threading, os, errno +import subprocess, threading, os, errno, time, datetime from Queue import Queue, Empty from datetime import datetime from galaxy import model # Database interaction class from galaxy.datatypes.data import nice_size -from time import sleep - import galaxy.eggs galaxy.eggs.require("boto") from boto.ec2.connection import EC2Connection @@ -19,64 +17,119 @@ """ Eucalyptus-based cloud provider implementation for managing instances. """ - def __init__( self, app, user ): + STOP_SIGNAL = object() + def __init__( self, app ): log.debug( "Using eucalyptus as default cloud provider." ) - self.conn = get_connection( user ) + self.zone = "epc" + self.key_pair = "galaxy-keypair" + + #TODO: Use multiple threads to process requests? - def get_connection( user ): + def get_connection( self, uci ): """ Establishes EC2 connection using user's default credentials """ log.debug( '##### Establishing cloud connection' ) - creds = model.CloudUserCredentials.filter_by( user=user, defaultCred=True ).first() - if creds: - a_key = creds.access_key - s_key = creds.secret_key - # Amazon EC2 - #conn = EC2Connection( a_key, s_key ) - # Eucalyptus Public Cloud - # TODO: Add option in Galaxy config file to specify these values (i.e., for locally manages Eucalyptus deployments) - euca_region = RegionInfo( None, "eucalyptus", "mayhem9.cs.ucsb.edu" ) - conn = EC2Connection( aws_access_key_id=a_key, aws_secret_access_key=s_key, is_secure=False, port=8773, region=euca_region, path="/services/Eucalyptus" ) - return conn - else: - log.debug( "User did not specify default credentials." ) - return 0 +# creds = model.CloudUserCredentials.filter_by( user=user, defaultCred=True ).first() + a_key = uci.credentials.access_key + s_key = uci.credentials.secret_key + # Amazon EC2 + #conn = EC2Connection( a_key, s_key ) + # Eucalyptus Public Cloud + # TODO: Add option in Galaxy config file to specify these values (i.e., for locally manages Eucalyptus deployments) + euca_region = RegionInfo( None, "eucalyptus", "mayhem9.cs.ucsb.edu" ) + conn = EC2Connection( aws_access_key_id=a_key, aws_secret_access_key=s_key, is_secure=False, port=8773, region=euca_region, path="/services/Eucalyptus" ) + return conn + + def get_keypair_name( self, uci, conn ): + """ + Generate keypair using user's default credentials + """ + log.debug( "Getting user's keypair" ) + kp = conn.get_key_pair( self.key_pair ) + + try: + for i, inst in enumerate( uci.instance ): + uci.instance[i].keypair_name = kp.name + return kp.name + except AttributeError: # No keypair under this name exists so create it + log.debug( "No keypair found, creating keypair '%s'" % self.key_pair ) + kp = conn.create_key_pair( self.key_pair ) + for i, inst in enumerate( uci.instance ): + uci.instance[i].keypair_name = kp.name + uci.instance[i].keypair_material = kp.material + uci.flush() + # TODO: Store key_pair.material into instance table - this is the only time private key can be retrieved + # Actually, probably return key_pair to calling method and store name & key from there... + + return kp.name + + def get_mi( self, type='small' ): + """ + Get appropriate machine image (mi) based on instance size. + TODO: Dummy method - need to implement logic + For valid sizes, see http://aws.amazon.com/ec2/instance-types/ + """ + return model.CloudImage.filter( model.CloudImage.table.c.id==1 ).first() + +# def get_instances( self, uci ): +# """ +# Get objects of instances that are pending or running and are connected to uci object +# """ +# instances = trans.sa_session.query( model.CloudInstance ) \ +# .filter_by( user=user, uci_id=uci.id ) \ +# .filter( or_(model.CloudInstance.table.c.state=="running", model.CloudInstance.table.c.state=="pending" ) ) \ +# .first() +# #.all() #TODO: return all but need to edit calling method(s) to handle list +# +# instances = uci.instance +# +# return instances - + def shutdown( self ): """Attempts to gracefully shut down the monitor thread""" log.info( "sending stop signal to worker threads in eucalyptus cloud manager" ) self.queue.put( self.STOP_SIGNAL ) log.info( "eucalyptus cloud manager stopped" ) - def createUCI( self, user, name, storage_size, zone=None): + def createUCI( self, uci ): """ Creates User Configured Instance (UCI). Essentially, creates storage volume on cloud provider and registers relevant information in Galaxy database. """ - conn = getConnection( user ) - # Capture user configured instance information - uci = model.UCI() - uci.name = name - uci.user = user - uci.state = "available" # Valid states include: "available", "running" or "pending" - uci.total_size = storage_size # This is OK now because a new instance is being created. - # Capture store related information - storage = model.CloudStore() - storage.user = user - storage.uci = uci - storage.size = storage_size - storage.availability_zone = "us-east-1a" # TODO: Give user choice here. Also, enable region selection. - #self.conn.create_volume( storage_size, storage.availability_zone, snapshot=None ) - # TODO: get correct value from Eucalyptus - storage.volume_id = "made up" - # Persist + conn = self.get_connection( uci ) + # Temporary code - need to ensure user selects zone at UCI creation time! + if uci.store[0].availability_zone=='': + log.info( "Availability zone for storage volume was not selected, using default zone: %s" % self.zone ) + uci.store[0].availability_zone = self.zone + uci.store[0].flush() + + #TODO: check if volume associated with UCI already exists (if server crashed for example) and don't recreate it + log.debug( "Creating volume in zone '%s'..." % uci.store[0].availability_zone ) + vol = conn.create_volume( uci.store[0].size, uci.store[0].availability_zone, snapshot=None ) + uci.store[0].volume_id = vol.id + + # Wait for a while to ensure volume was created +# vol_status = vol.status +# for i in range( 30 ): +# if vol_status is not "u'available'": +# log.debug( 'Updating volume status; current status: %s' % vol_status ) +# vol_status = vol.status +# time.sleep(3) +# if i is 29: +# log.debug( "Error while creating volume '%s'; stuck in state '%s'; deleting volume." % ( vol.id, vol_status ) ) +# conn.delete_volume( vol.id ) +# uci.state = 'error' +# uci.flush() +# return + + uci.state = 'available' + uci.store[0].status = vol.status + uci.store[0].flush() uci.flush() - storage.flush() - session.flush() - + def deleteUCI( self, name ): """ Deletes UCI. NOTE that this implies deletion of any and all data associated @@ -86,14 +139,142 @@ def addStorageToUCI( self, name ): """ Adds more storage to specified UCI """ - def startUCI( self, name, type ): + def startUCI( self, uci ): """ Starts an instance of named UCI on the cloud. This implies, mounting of storage and starting Galaxy instance. """ + conn = self.get_connection( uci ) - def stopUCI( self, name ): + uci.instance[0].keypair_name = self.get_keypair_name( uci, conn ) + mi = self.get_mi( uci.instance[0].type ) + +# log.debug( "mi: %s, mi.image_id: %s, uci.instance[0].keypair_name: %s" % ( mi, mi.image_id, uci.instance[0].keypair_name ) ) + uci.instance[0].image = mi + +# log.debug( '***** Setting up security group' ) + # If not existent, setup galaxy security group +# try: +# gSecurityGroup = conn.create_security_group('galaxy', 'Security group for Galaxy.') +# gSecurityGroup.authorize( 'tcp', 80, 80, '0.0.0.0/0' ) # Open HTTP port +# gSecurityGroup.authorize( 'tcp', 22, 22, '0.0.0.0/0' ) # Open SSH port +# except: +# pass +# sgs = conn.get_all_security_groups() +# for i in range( len( sgs ) ): +# if sgs[i].name == "galaxy": +# sg.append( sgs[i] ) +# break # only 1 security group w/ this name can exist, so continue + + log.debug( "***** Starting UCI instance '%s'" % uci.name ) +# log.debug( 'Using following command: conn.run_instances( image_id=%s, key_name=%s )' % ( uci.instance[0].image.image_id, uci.instance[0].keypair_name ) ) + reservation = conn.run_instances( image_id=uci.instance[0].image.image_id, key_name=uci.instance[0].keypair_name ) + #reservation = conn.run_instances( image_id=instance.image, key_name=instance.keypair_name, security_groups=['galaxy'], instance_type=instance.type, placement=instance.availability_zone ) + uci.instance[0].launch_time = datetime.utcnow() + uci.launch_time = uci.instance[0].launch_time + uci.instance[0].reservation_id = str( reservation ).split(":")[1] + uci.instance[0].instance_id = str( reservation.instances[0]).split(":")[1] + s = reservation.instances[0].state + uci.instance[0].state = s + uci.state = s + uci.instance[0].flush() + uci.flush() + + # Wait until instance gets running and then update the DB + while s!="running": + log.debug( "Waiting on instance '%s' to start up (reservation ID: %s); current state: %s" % ( uci.instance[0].instance_id, uci.instance[0].reservation_id, s ) ) + time.sleep( 15 ) + s = reservation.instances[0].update() + + uci.instance[0].state = s + uci.state = s + uci.instance[0].public_dns = reservation.instances[0].dns_name + uci.instance[0].private_dns = reservation.instances[0].private_dns_name + uci.instance[0].flush() + uci.flush() + + + def stopUCI( self, uci ): """ - Stops cloud instance associated with named UCI. This also implies - stopping of Galaxy and unmounting of the file system. - """ \ No newline at end of file + Stops all of cloud instances associated with named UCI. + """ + conn = self.get_connection( uci ) + tl = [] # temination list + + for i, inst in enumerate( uci.instance ): + tl.append( uci.instance[i].instance_id ) + + instList = conn.get_all_instances( tl ) +# log.debug( 'instList: %s' % instList ) + + for i, inst in enumerate( instList ): +# log.debug( 'inst: %s' % inst ) + log.debug( 'Before stop - inst.instances[0].update(): %s' % inst.instances[0].update() ) + inst.instances[0].stop() + log.debug( 'After stop - inst.instances[0].update(): %s' % inst.instances[0].update() ) + uci.instance[i].stop_time = datetime.utcnow() + + terminated=0 + while terminated!=len( instList ): + for i, inst in enumerate( instList ): + log.debug( "inst state: %s" % inst.instances[0].state ) + state = inst.instances[0].update() + if state=='terminated': + uci.instance[i].state = state + uci.instance[i].flush() + terminated += 1 + time.sleep ( 5 ) + + uci.state = 'available' + uci.launch_time = None + uci.flush() + + log.debug( "All instances for UCI '%s' were terminated." % uci.name ) + + + +# dbInstances = get_instances( trans, uci ) #TODO: handle list! +# +# # Get actual cloud instance object +# cloudInstance = get_cloud_instance( conn, dbInstances.instance_id ) +# +# # TODO: Detach persistent storage volume(s) from instance and update volume data in local database +# stores = get_stores( trans, uci ) +# for i, store in enumerate( stores ): +# log.debug( "Detaching volume '%s' to instance '%s'." % ( store.volume_id, dbInstances.instance_id ) ) +# mntDevice = store.device +# volStat = None +## Detaching volume does not work with Eucalyptus Public Cloud, so comment it out +## try: +## volStat = conn.detach_volume( store.volume_id, dbInstances.instance_id, mntDevice ) +## except: +## log.debug ( 'Error detaching volume; still going to try and stop instance %s.' % dbInstances.instance_id ) +# store.attach_time = None +# store.device = None +# store.i_id = None +# store.status = volStat +# log.debug ( '***** volume status: %s' % volStat ) +# +# +# # Stop the instance and update status in local database +# cloudInstance.stop() +# dbInstances.stop_time = datetime.utcnow() +# while cloudInstance.state != 'terminated': +# log.debug( "Stopping instance %s state; current state: %s" % ( str( cloudInstance ).split(":")[1], cloudInstance.state ) ) +# time.sleep(3) +# cloudInstance.update() +# dbInstances.state = cloudInstance.state +# +# # Reset relevant UCI fields +# uci.state = 'available' +# uci.launch_time = None +# +# # Persist +# session = trans.sa_session +## session.save_or_update( stores ) +# session.save_or_update( dbInstances ) # TODO: Is this going to work w/ multiple instances stored in dbInstances variable? +# session.save_or_update( uci ) +# session.flush() +# trans.log_event( "User stopped cloud instance '%s'" % uci.name ) +# trans.set_message( "Galaxy instance '%s' stopped." % uci.name ) + \ No newline at end of file diff -r c9c9adf06e9d -r 7c438fd3cf4a lib/galaxy/model/mapping.py --- a/lib/galaxy/model/mapping.py Wed Sep 30 17:57:11 2009 -0400 +++ b/lib/galaxy/model/mapping.py Fri Oct 02 18:31:32 2009 -0400 @@ -432,9 +432,9 @@ Column( "attach_time", DateTime ), Column( "user_id", Integer, ForeignKey( "galaxy_user.id" ), index=True, nullable=False ), Column( "uci_id", Integer, ForeignKey( "uci.id" ), index=True, nullable=False ), - Column( "volume_id", TEXT, nullable=False ), + Column( "volume_id", TEXT ), Column( "size", Integer, nullable=False ), - Column( "availability_zone", TEXT, nullable=False ), + Column( "availability_zone", TEXT ), Column( "i_id", TEXT, ForeignKey( "cloud_instance.instance_id" ), index=True ), Column( "status", TEXT ), Column( "device", TEXT ), diff -r c9c9adf06e9d -r 7c438fd3cf4a lib/galaxy/model/migrate/versions/0014_cloud_tables.py --- a/lib/galaxy/model/migrate/versions/0014_cloud_tables.py Wed Sep 30 17:57:11 2009 -0400 +++ b/lib/galaxy/model/migrate/versions/0014_cloud_tables.py Fri Oct 02 18:31:32 2009 -0400 @@ -38,11 +38,11 @@ Column( "launch_time", DateTime ), Column( "stop_time", DateTime ), Column( "user_id", Integer, ForeignKey( "galaxy_user.id" ), index=True, nullable=False ), - Column( "uci_id", Integer, ForeignKey( "uci.id" ), index=True ), + Column( "uci_id", Integer, ForeignKey( "uci.id" ), index=True, nullable=False ), Column( "type", TEXT ), Column( "reservation_id", TEXT ), Column( "instance_id", TEXT ), - Column( "mi_id", TEXT, ForeignKey( "cloud_image.image_id" ), index=True, nullable=False ), + Column( "mi_id", TEXT, ForeignKey( "cloud_image.image_id" ), index=True ), Column( "state", TEXT ), Column( "public_dns", TEXT ), Column( "private_dns", TEXT ), @@ -57,9 +57,9 @@ Column( "attach_time", DateTime ), Column( "user_id", Integer, ForeignKey( "galaxy_user.id" ), index=True, nullable=False ), Column( "uci_id", Integer, ForeignKey( "uci.id" ), index=True, nullable=False ), - Column( "volume_id", TEXT, nullable=False ), + Column( "volume_id", TEXT ), Column( "size", Integer, nullable=False ), - Column( "availability_zone", TEXT, nullable=False ), + Column( "availability_zone", TEXT ), Column( "i_id", TEXT, ForeignKey( "cloud_instance.instance_id" ), index=True ), Column( "status", TEXT ), Column( "device", TEXT ), @@ -93,8 +93,8 @@ def downgrade(): metadata.reflect() try: -# log.debug( "Would drop cloud_image table." ) - CloudImage_table.drop() #Enable before release + log.debug( "Would drop cloud_image table." ) +# CloudImage_table.drop() #Enable before release except Exception, e: log.debug( "Dropping cloud_image table failed: %s" % str( e ) ) @@ -109,8 +109,8 @@ log.debug( "Dropping cloud_store table failed: %s" % str( e ) ) try: -# log.debug( "Would drop cloud_user_credentials table." ) - CloudUserCredentials_table.drop() #Enable before putting final version + log.debug( "Would drop cloud_user_credentials table." ) +# CloudUserCredentials_table.drop() #Enable before putting final version except Exception, e: log.debug( "Dropping cloud_user_credentials table failed: %s" % str( e ) ) diff -r c9c9adf06e9d -r 7c438fd3cf4a lib/galaxy/web/controllers/cloud.py --- a/lib/galaxy/web/controllers/cloud.py Wed Sep 30 17:57:11 2009 -0400 +++ b/lib/galaxy/web/controllers/cloud.py Fri Oct 02 18:31:32 2009 -0400 @@ -42,12 +42,12 @@ Render cloud main page (management of cloud resources) """ user = trans.get_user() - pendingInstances = trans.sa_session.query( model.UCI ) \ - .filter_by( user=user, state="pending" ) \ - .all() - - for i in range( len ( pendingInstances ) ): - update_instance_state( trans, pendingInstances[i].id ) +# pendingInstances = trans.sa_session.query( model.UCI ) \ +# .filter_by( user=user, state="pending" ) \ +# .all() +# +# for i inupdate_in range( len ( pendingInstances ) ): +# stance_state( trans, pendingInstances[i].id ) cloudCredentials = trans.sa_session.query( model.CloudUserCredentials ) \ .filter_by( user=user ) \ @@ -56,12 +56,13 @@ liveInstances = trans.sa_session.query( model.UCI ) \ .filter_by( user=user ) \ - .filter( or_(model.UCI.c.state=="running", model.UCI.c.state=="pending") ) \ + .filter( or_( model.UCI.c.state=="running", model.UCI.c.state=="pending", model.UCI.c.state=="terminating" ) ) \ .order_by( desc( model.UCI.c.launch_time ) ) \ .all() prevInstances = trans.sa_session.query( model.UCI ) \ - .filter_by( user=user, state="available" ) \ + .filter_by( user=user ) \ + .filter( or_( model.UCI.c.state=="available", model.UCI.c.state=="new", model.UCI.c.state=="error", model.UCI.c.state=="submitted" ) ) \ .order_by( desc( model.UCI.c.update_time ) ) \ .all() @@ -101,26 +102,26 @@ @web.expose @web.require_login( "start Galaxy cloud instance" ) - def start( self, trans, id, size='small' ): + def start( self, trans, id, type='small' ): """ Start a new cloud resource instance """ - # TODO: Add choice of instance size before starting one - #if size: + # TODO: Add choice of instance type before starting one + #if type: user = trans.get_user() - mi = get_mi( trans, size ) + mi = get_mi( trans, type ) uci = get_uci( trans, id ) stores = get_stores( trans, uci ) - log.debug(self.app.config.job_working_directory) +# log.debug(self.app.config.job_working_directory) if len(stores) is not 0: instance = model.CloudInstance() instance.user = user instance.image = mi instance.uci = uci - instance.keypair_name = get_keypair_name( trans ) instance.availability_zone = stores[0].availability_zone # Bc. all EBS volumes need to be in the same avail. zone, just check 1st - instance.type = size - conn = get_connection( trans ) + instance.type = type +# instance.keypair_name = get_keypair_name( trans ) +# conn = get_connection( trans ) # log.debug( '***** Setting up security group' ) # If not existent, setup galaxy security group # try: @@ -135,26 +136,25 @@ # sg.append( sgs[i] ) # break # only 1 security group w/ this name can exist, so continue - log.debug( '***** Starting an instance' ) - log.debug( 'Using following command: conn.run_instances( image_id=%s, key_name=%s )' % ( instance.image.image_id, instance.keypair_name ) ) - reservation = conn.run_instances( image_id=instance.image.image_id, key_name=instance.keypair_name ) +# log.debug( '***** Starting an instance' ) +# log.debug( 'Using following command: conn.run_instances( image_id=%s, key_name=%s )' % ( instance.image.image_id, instance.keypair_name ) ) +# reservation = conn.run_instances( image_id=instance.image.image_id, key_name=instance.keypair_name ) #reservation = conn.run_instances( image_id=instance.image, key_name=instance.keypair_name, security_groups=['galaxy'], instance_type=instance.type, placement=instance.availability_zone ) - instance.launch_time = datetime.utcnow() - uci.launch_time = instance.launch_time - instance.reservation_id = str( reservation ).split(":")[1] - instance.instance_id = str( reservation.instances[0]).split(":")[1] +# instance.launch_time = datetime.utcnow() +# uci.launch_time = instance.launch_time +# instance.reservation_id = str( reservation ).split(":")[1] +# instance.instance_id = str( reservation.instances[0]).split(":")[1] # instance.state = "pending" - instance.state = reservation.instances[0].state - uci.state = instance.state - - # TODO: After instance boots up, need to update status, DNS and attach EBS +# instance.state = reservation.instances[0].state + uci.state = 'submitted' # Persist session = trans.sa_session session.save_or_update( instance ) + session.save_or_update( uci ) session.flush() - trans.log_event ("Started new instance. Reservation ID: '%s', Instance ID: '%s'" % (instance.reservation_id, instance.instance_id ) ) + trans.log_event ("User initiated starting of cloud instance '%s'." % uci.name ) return self.list( trans ) # return trans.show_form( @@ -167,55 +167,63 @@ @web.require_login( "stop Galaxy cloud instance" ) def stop( self, trans, id ): """ - Stop a cloud instance. This implies stopping Galaxy server and disconnecting/unmounting relevant file system(s). + Stop a cloud UCI instance. This implies stopping Galaxy server and disconnecting/unmounting relevant file system(s). """ uci = get_uci( trans, id ) - dbInstances = get_instances( trans, uci ) #TODO: handle list! - - conn = get_connection( trans ) - # Get actual cloud instance object - cloudInstance = get_cloud_instance( conn, dbInstances.instance_id ) - - # TODO: Detach persistent storage volume(s) from instance and update volume data in local database - stores = get_stores( trans, uci ) - for i, store in enumerate( stores ): - log.debug( "Detaching volume '%s' to instance '%s'." % ( store.volume_id, dbInstances.instance_id ) ) - mntDevice = store.device - volStat = None -# Detaching volume does not work with Eucalyptus Public Cloud, so comment it out -# try: -# volStat = conn.detach_volume( store.volume_id, dbInstances.instance_id, mntDevice ) -# except: -# log.debug ( 'Error detaching volume; still going to try and stop instance %s.' % dbInstances.instance_id ) - store.attach_time = None - store.device = None - store.i_id = None - store.status = volStat - log.debug ( '***** volume status: %s' % volStat ) - - - # Stop the instance and update status in local database - cloudInstance.stop() - dbInstances.stop_time = datetime.utcnow() - while cloudInstance.state != 'terminated': - log.debug( "Stopping instance %s state; current state: %s" % ( str( cloudInstance ).split(":")[1], cloudInstance.state ) ) - time.sleep(3) - cloudInstance.update() - dbInstances.state = cloudInstance.state - - # Reset relevant UCI fields - uci.state = 'available' - uci.launch_time = None - - # Persist + uci.state = 'terminating' session = trans.sa_session # session.save_or_update( stores ) - session.save_or_update( dbInstances ) # TODO: Is this going to work w/ multiple instances stored in dbInstances variable? session.save_or_update( uci ) session.flush() trans.log_event( "User stopped cloud instance '%s'" % uci.name ) trans.set_message( "Galaxy instance '%s' stopped." % uci.name ) - + +# dbInstances = get_instances( trans, uci ) #TODO: handle list! +# +# conn = get_connection( trans ) +# # Get actual cloud instance object +# cloudInstance = get_cloud_instance( conn, dbInstances.instance_id ) +# +# # TODO: Detach persistent storage volume(s) from instance and update volume data in local database +# stores = get_stores( trans, uci ) +# for i, store in enumerate( stores ): +# log.debug( "Detaching volume '%s' to instance '%s'." % ( store.volume_id, dbInstances.instance_id ) ) +# mntDevice = store.device +# volStat = None +## Detaching volume does not work with Eucalyptus Public Cloud, so comment it out +## try: +## volStat = conn.detach_volume( store.volume_id, dbInstances.instance_id, mntDevice ) +## except: +## log.debug ( 'Error detaching volume; still going to try and stop instance %s.' % dbInstances.instance_id ) +# store.attach_time = None +# store.device = None +# store.i_id = None +# store.status = volStat +# log.debug ( '***** volume status: %s' % volStat ) +# +# +# # Stop the instance and update status in local database +# cloudInstance.stop() +# dbInstances.stop_time = datetime.utcnow() +# while cloudInstance.state != 'terminated': +# log.debug( "Stopping instance %s state; current state: %s" % ( str( cloudInstance ).split(":")[1], cloudInstance.state ) ) +# time.sleep(3) +# cloudInstance.update() +# dbInstances.state = cloudInstance.state +# +# # Reset relevant UCI fields +# uci.state = 'available' +# uci.launch_time = None +# +# # Persist +# session = trans.sa_session +## session.save_or_update( stores ) +# session.save_or_update( dbInstances ) # TODO: Is this going to work w/ multiple instances stored in dbInstances variable? +# session.save_or_update( uci ) +# session.flush() +# trans.log_event( "User stopped cloud instance '%s'" % uci.name ) +# trans.set_message( "Galaxy instance '%s' stopped." % uci.name ) +# return self.list( trans ) @web.expose @@ -302,20 +310,12 @@ trans.app.model.CloudUserCredentials.table.c.name==credName ).first() uci.user= user uci.total_size = volSize # This is OK now because new instance is being created. - # Need to flush because connection object accesses uci table - uci.flush() - # Capture store related information + uci.state = "new" # Valid states include: "new, "available", "running", "pending", "submitted", "terminating", or "error" storage = model.CloudStore() storage.user = user storage.uci = uci storage.size = volSize - storage.availability_zone = "us-east-1a" # TODO: Give user choice here. Also, enable region selection. - conn = get_connection( trans, credName ) - #conn.create_volume( volSize, storage.availability_zone, snapshot=None ) - # TODO: get correct value from AWS - storage.volume_id = "made up" - # TODO: If volume creation was successfull, set state to available - uci.state = "available" # Valid states include: "available", "running" or "pending" + storage.availability_zone = zone # TODO: Give user choice here. Also, enable region selection. # Persist session = trans.sa_session session.save_or_update( uci ) @@ -323,7 +323,7 @@ session.flush() # Log and display the management page trans.log_event( "User configured new cloud instance" ) - trans.set_message( "New Galaxy instance '%s' configured." % instanceName ) + trans.set_message( "New Galaxy instance '%s' configured. Once instance status shows 'available' you will be able to start the instance." % instanceName ) return self.list( trans ) except ValueError: vol_error = "Volume size must be specified as an integer value only, between 1 and 1000."

1 0