Hi Dev Team.
We've been running a local galaxy instance on an OS X Xserve for the past
several years. You guys have been very helpful with problems we've had in
the past and I'm hoping you can point us in the right direction for our
current suite of server errors.
We upgraded to the latest version with Hg over a week ago, and did not
notice any problems, either with our custom tools or with the API scripting
that we use to recursively execute workflows.
However, for reasons we do not yet understand, yesterday the server started
throwing errors at us, most of which seem to be related to our galaxy
postgres database. After many hours of troubleshooting we were able to
restore our local server to a point where we can download previously
created datasets and export our workflows (which was a huge relief). My PI
(Dr. David O'Connor who is cc'd here) did most of the troubleshooting to
get us to that point and can provide greater details on that if needed.
However, we still cannot execute tools on any datasets. We cannot upload
new datasets, however we can pull them into the current history from
previously created data libraries.
We feel there may be duplication or corruption in the postgres database. It
appears that the Galaxy can pull previous data from the database, but that
sqlalchemy cannot insert properly. Is there any way you know of to cleanup
our database if this is the case? I see the manage_db.py script in the
server directory, but do not want to do anything that will further
complicate the problem.
To check if there were dependency issues (as opposed to database problems)
we downloaded and installed a clean server version, which is working fine
with the built-in sql-light database.
We are basically looking for advice on: 1) If this is a database issue, can
it be fixed using any of Galaxy's scripts or tools? 2) What other tests can
we try to determine if this is not a database problem? and 3) Is it
possible to migrate our existing database to a clean install if there are
deeper structural problems with our local instance.
I don't want to send you guys all of our errors since I am sure you are
busy (unless that would be helpful and you could take the time to look at
them), but I've included a representative error that occurs when trying to
execute a tool below.
Thanks so much for your help in the past and any advice you might have.
Simon Lank
Research Specialist
O'Connor Lab, WNPRC
555 Science Dr. Madison WI
(608) 265-3389
Error executing tool: (ProgrammingError) column "params" of relation "job"
does not exist LINE 1: ...ob_runner_external_id, object_store_id, imported,
params) VA... ^ 'INSERT INTO job (id, create_time, update_time, history_id,
library_folder_id, tool_id, tool_version, state, info, command_line,
param_filename, runner_name, stdout, stderr, traceback, session_id,
user_id, job_runner_name, job_runner_external_id, object_store_id,
imported, params) VALUES (%(id)s, %(create_time)s, %(update_time)s,
%(history_id)s, %(library_folder_id)s, %(tool_id)s, %(tool_version)s,
%(state)s, %(info)s, %(command_line)s, %(param_filename)s, %(runner_name)s,
%(stdout)s, %(stderr)s, %(traceback)s, %(session_id)s, %(user_id)s,
%(job_runner_name)s, %(job_runner_external_id)s, %(object_store_id)s,
%(imported)s, %(params)s)' {'stdout': None, 'create_time':
datetime.datetime(2012, 4, 17, 16, 11, 22, 676616), 'id': 142237L,
'runner_name': None, 'user_id': 3, 'tool_version': '1.0.0', 'command_line':
None, 'state': 'new', 'params': None, 'job_runner_name': None, 'tool_id':
'Show beginning1', 'update_time': datetime.datetime(2012, 4, 17, 16, 11,
22, 676632), 'history_id': 2330, 'param_filename': None, 'imported': False,
'object_store_id': None, 'info': None, 'traceback': None, 'session_id':
3272, 'library_folder_id': None, 'stderr': None, 'job_runner_external_id':
None}