September 2015 - galaxy-dev - lists.galaxyproject.org

Cloudman - using different worker instance types
by David Kovalic 13 Sep '15

13 Sep '15

Hi, I want to be able to use different AWS worker instance types than are currently provided in the cloudman "Add worker nodes" menu. Say for example I am interested in using AWS 'r3.*' instance types as my workers, how can I add these to be available in the cloudman interface? Thanks, David

2 2

Adding categories to Main
by Keith Suderman 12 Sep '15

12 Sep '15

Dear Galaxy Team, A colleague would like to upload some NLP (Natural Language Processing) tools to the Galaxy Test/Main tools sheds, but we are unclear what "categories" to use for the tools. I see that the Main/Test tools sheds have a category for "Text Manipulation", but that does not seem appropriate for NLP tools. Is it possible to have new categories added to the tool shed(s)? If so, what is the process? I am just starting to investigate setting up our own local tool shed and I am coming across mentions of repository capsules and exporting tool sheds. Would it be preferable to install the NLP tools to a local tool shed and then export a repository/capsule to be imported to the Test/Main tools sheds? What happens if our tool shed uses a disjoint set of categories than Test/Main? Cheers, Keith ------------------------------ Research Associate Department of Computer Science Vassar College Poughkeepsie, NY

4 9

custom tool - accessing user's data in bash script
by Piotr Grabowski 11 Sep '15

11 Sep '15

Hello Galaxy Devs and Deployers! I have a rather basic question about how to handle custom tools in Galaxy (I'm fresh to this framework). I am trying to link my local Galaxy with KNIME workflow that we developed in our group. I am trying to link it via a bash script that is running headless KNIME and it seems almost to work, but... In order for this workflow to run I have to specify in the command line things like result folder or where a dynamically generated parameter file is located. This parameter file is generate by the bash script before running KNIME. The paths will be always dynamic, depending on users etc. So I can't simply hard-code the path. Is there a way to make bash aware of where the current Galaxy user's folders and files are upon execution in Galaxy? Below a simplified scheme of what I am doing: User uploads txt file -> Clicks RUN -> bash script builds a parameter file (gawk ... >> param.csv) -> /knime run workflow --set the path to the parameter .csv file + set export folder Any ideas ? Best, Piotr

2 1

LDAP and valid username
by Edouard Hirchaud 11 Sep '15

11 Sep '15

Hi, I configure authentication with the LDAP (Active Directory) of my compagny. But the username (sAMAccoutName) contains a dot (eg : e.hirchaud). And that was a problem because the dot is not allow in galaxy username and auto-registration fails because of that. For testing I change the regex in "security/validate_user_input.py" and authorize character dot and it's work ! I know its not a good practice to change core code so I want to know if they are known "risks" or sides effects if I keep this change and if you have a best solution ? Thanks! Edouard -- Édouard Hirchaud ANSES Laboratoire de Ploufragan-Plouzané Unité Génétique Virale et Biosécurité 22440 Ploufragan Tel : (33) 02 96 01 62 71

1 0

Tool with multiple outputs.
by Léo Biscassi 11 Sep '15

11 Sep '15

Good afternoon people! I am developing a tool that will have multiple outputs where they have to be data sets in history. Read the tutorial [1] and from it developed my [2] [3] tool, however, have only the .txt output file, the .mol2 files are not showing up in history, someone can tell me why? [1] https://wiki.galaxyproject.org/Admin/Tools/Multiple%20Output%20Files#Handli… [2] https://gist.github.com/leobiscassi/8058b6307d388167866a [3] https://gist.github.com/leobiscassi/bc54a7df261c4ee81aa9 Att. -- *Léo Biscassi* Facebook - https://pt-br.facebook.com/lbiscassi Linkedin - https://br.linkedin.com/in/leobiscassi Skype - biscassi.leo Twitter - https://twitter.com/leobiscassi -

3 4

Fatal Error while uploading a File with special Characters
by Matthias Enders 11 Sep '15

11 Sep '15

Dear Galaxy Community, a User of our Galaxy Platform has produced a fatal error by uploading a file containing a German special character (ä / ü / ö / ß). This resulted in the following status (only for the given user): - Upload was canceled: Status was "Failed Internal Server Error (500)" - All Histories were corrupt stating: "An error was encountered while loading the contents." - Tools can´t be used an stating (in Traceback): " UnicodeDecodeError: 'utf8' codec can't decode byte 0xfc in position 5: invalid start byte" "Module weberror.evalexception.middleware:364 in respond" This totally destroyed to current user and we have had to purge him in order to get him back to work, even losing all his datasets & histories! We are using an MySQL database as a backend. Is the Encoding of this database the problem? If yes: Which encoding is used/supported by galaxy? Thanks in advance! Matthias

2 4

Configuring Apache server with uwsgi
by Makis Ladoukakis 10 Sep '15

10 Sep '15

Hi everyone, I am setting up a galaxy instance in my university server that has 64 cores (4 cpus 16 threads each) and 512GB of memory and I am trying to figure out the best configuration in order to exploit it as best as we can. We are planning to use it for about 100-200 people. So far I followed the instructions in https://wiki.galaxyproject.org/Admin/GetGalaxy https://wiki.galaxyproject.org/Admin/Config/Performance/ProductionServer https://wiki.galaxyproject.org/Admin/Config/ApacheProxy and https://wiki.galaxyproject.org/Admin/Config/Performance/Scaling but I found a little bit confusing the different instructions for Paste based processes and uwsgi. I'm planning to use uwsgi in Apache. So I've been using mysql as the galaxy database and added the following part in galaxy.ini [uwsgi] processes = 12 stats = 127.0.0.1:9191 socket = 127.0.0.1:4001 pythonpath = lib threads = 4 logto = /home/galaxy/uwsgi.log master = True [server:handler0] use = egg:Paste#http port = 8090 use_threadpool = True threadpool_workers = 10 [server:handler1] use = egg:Paste#http port = 8091 use_threadpool = True threadpool_workers = 10 [server:handler2] use = egg:Paste#http port = 8092 use_threadpool = True threadpool_workers = 10 [server:handler3] use = egg:Paste#http port = 8093 use_threadpool = True threadpool_workers = 10 [server:handler4] use = egg:Paste#http port = 8094 use_threadpool = True threadpool_workers = 10 [server:handler5] use = egg:Paste#http port = 8095 use_threadpool = True threadpool_workers = 10 # The port on which to listen. #port = 8090 >From a previous post i got that you are using 6 handlers for usegalaxy.org so I figured I'd start with that. Could you help me configure my apache server? So far I've added the following in the apache configuration file: RewriteEngine on RewriteRule ^(.*) http://localhost:8090$1 [P] RewriteRule ^/galaxy$ /galaxy/ [R] RewriteRule ^/galaxy/static/style/(.*) /home/nate/galaxy-dist/static/june_2007_style/blue/$1 [L] RewriteRule ^/galaxy/static/scripts/(.*) /home/nate/galaxy-dist/static/scripts/packed/$1 [L] RewriteRule ^/galaxy/static/(.*) /home/nate/galaxy-dist/static/$1 [L] RewriteRule ^/galaxy/favicon.ico /home/nate/galaxy-dist/static/favicon.ico [L] RewriteRule ^/galaxy/robots.txt /home/nate/galaxy-dist/static/robots.txt [L] <Location "/galaxy"> Sethandler uwsgi-handler uWSGISocket 127.0.0.1:4001 uWSGImaxVars 512 </Location> But I'm not sure how Apache will proxy the different ports from the handler. In the Apache configuration instructions it said I should use a balancer but in the Scaling instructions it states that this is for standalone paste-based processes. Could you help me out please? Thank you, Makis

1 0

galaxy.ini tool_dependency_dir change
by Dooley, Damion 09 Sep '15

09 Sep '15

On a VM I moved the tool_dependency_dir folder to another mount since it was getting so large that the smallish main mount was full (and changed galaxy.ini, restarted galaxy, and checked that permissions were appropriate for galaxy user). But I'm finding I have to reinstall various tool dependencies (prokka etc) to enable galaxy to find the command line executables. Is there a trick to getting the tool dependencies working again after moving the folder wholesale? Thanks, Damion Hsiao lab, BC Public Health Microbiology & Reference Laboratory, BC Centre for Disease Control 655 West 12th Avenue, Vancouver, British Columbia, V5Z 4R4 Canada

3 4

BUG: Collections and History export
by Alexander Vowinkel 09 Sep '15

09 Sep '15

Hi, when downloading/exporting a history via the interface, visible collections are not being exported. The contents of the collection(s) are hidden, the Colleciton itself is visible. Best, Alexander

2 1

Connection to PBS server XXX for state check failed
by Zuzanna K. Filutowska 09 Sep '15

09 Sep '15

Dear All, I encountered exactly the same problem: http://osdir.lowified.com/galaxy-development-source-control/2012-07/msg0013… However my Galaxy instance completely crashes until restart, I am getting Proxy Error on HTTPd. Any help would be much appreciated, as we will have Galaxy workshop on Monday at our University. In logs I have: galaxy.jobs.runners.pbs DEBUG 2015-09-09 15:04:21,306 connection to PBS server ce.reef.man.poznan.pl for state check failed galaxy.jobs.runners.pbs DEBUG 2015-09-09 15:04:21,380 (441/13320109.ce.reef.man.poznan.pl) Skipping state check because PBS server connection failed galaxy.jobs.runners.pbs DEBUG 2015-09-09 15:04:21,430 (444/13320114.ce.reef.man.poznan.pl) Skipping state check because PBS server connection failed +2+12+19+6galaxy+0+32+12+9job_state+1+0+0+02+172+14resources_used+1+0+0 +02+142+11exit_status+1+0+0+0+0 But the PBS server works fine. -- Pozdrawiam, -- Zuzanna K. Filutowska, HPC Systems Administrator Poznan Supercomputing and Networking Center Institute of Bioorganic Chemistry Polish Academy of Sciences Seize the day boys! Make your lifes extraordinary! --John Keating

2 2