- galaxy-dev - lists.galaxyproject.org

move dataset from history to server
by colin molter 30 Jul '11

30 Jul '11

Hello, Is there a way to copy/move a dataset located in the history to a given location in the filesystem? As an administrator you can upload data from given directory to your datalibrary. Is there a way to do the opposite? I know that it is possible to download any dataset to your local computer but what I want is just to move the dataset constructed by galaxy to a given repository in my server filesystem. thank you, colin

1 0

exporting environment variables to SGE in galaxy
by Chaolin Zhang 30 Jul '11

30 Jul '11

Hi, I am setting up SGE in our galaxy mirror. One problem I have is that I cannot export environment variables of the specific users running the galaxy service. On command line, I did this by qsub -V script.sh. or add a line #$ -V in script.sh I tried to change /lib/galaxy/jobs/runners/sge.py and lib/galaxy/jobs/runners/drmaae.py by adding a line #$ -V under sge_template = """#!/bin/sh #$ -S /bin/sh but this did not help. Any idea is very appreciated! Chaolin

2 2

"Job output not returned from cluster"
by Edward Kirton 30 Jul '11

30 Jul '11

I've been getting these errors sometimes lately, particularly when the cluster is heavily loaded. The jobs have completed successfully, as I can see the output if I click the pen icon, but the job is in a failed state. Have any other sites been experiencing this problem? Or can the galaxy developers help shed some light on the issue? FYI, I use the outputs_to_working_directory option in universe_wsgi.ini so that i can use a faster/more reliable filesystem to collect output from the cluster. I'm not using the recently discussed patch to run jobs as the unix user. I'll continue to experiment with different filesystems and software settings.

3 4

BWA sampe runs 'forever'
by Assaf Gordon 30 Jul '11

30 Jul '11

Hello, Question for galaxy maintainers: have you encountered situations where BWA jobs run 'forever' (for days) ? A little digging shows that it's the "bwa sampe" step, and SEQanswers thread mention it's somewhat common: http://seqanswers.com/forums/showthread.php?t=11652 http://seqanswers.com/forums/showthread.php?t=6340 One suggested work-around is to add "-A" to "bwa sampe", but there's no way to do it in the current BWA wrapper. (On the command-line with the same input files it does make things much faster). I'm wondering if this is common enough to justify adding "-A" to the wrapper, or is it rare enough and can be ignored. -gordon

2 3

manage library permission interface glitch
by Glen Beane 30 Jul '11

30 Jul '11

I should preface this by saying we are running a galaxy-dist that is a couple months old now (we are planning on updating soon), but here is an annoying issue we are seeing managing data library permissions: The Roles not associated lists are filtered so it shows only roles associated with the access library permission -- this is good, except this is filtered for the access library permission too, meaning once you add roles to the access library permission and save you can't add any other roles to this permission since the "Roles not associated" for access library now only includes roles associated with access library. To add any new roles we need to first remove every role from this permission, making it public, which then causes all roles to get listed in "roles not associated", and then we can associate all of the roles we want. The correct behavior would be to filter every "roles not associated" list for every permission except for "access library" -- this should always list all roles not associated. I will open a new issue for this on Bitbucket unless a developer can confirm that it has been fixed since my last galaxy-dist update a couple months ago. -- Glen L. Beane Senior Software Engineer The Jackson Laboratory (207) 288-6153

2 2

problem with job runner
by shashi shekhar 29 Jul '11

29 Jul '11

Hi In my galaxy instance, whatever jobs i am submitting it goes into queued state. If I restart the server then the previous submitted jobs state changes to running. but the newly submitted jobs again goes to queued state. I am at a loss to understand this behaviour of galaxy and unable to debug it. The job submission uses a customized runner. How is it actually goes into the queued state automatically when all the workers thread are free? Does galaxy_session table is_valid attribute makes jobs state true? Or what all places in tables the queued states are getting stored. I can only see that the jobs table state attribute only stores the state. The server logs points error here: galaxy.jobs ERROR 2011-07-29 11:01:28,098 failure running job 2243 Traceback (most recent call last): File "/home/gwadmin/galaxy-central/ lib/galaxy/jobs/__init__.py", line 202, in __monitor_step self.dispatcher.put( JobWrapper( job, self ) ) File "/home/gwadmin/galaxy-central/lib/galaxy/jobs/__init__.py", line 856, in put self.job_runners[runner_name].put( job_wrapper ) File "/home/gwadmin/galaxy-central/lib/galaxy/jobs/runners/gw.py", line 375, in put job_wrapper.change_state( model.Job.states.QUEUED ) File "/home/gwadmin/galaxy-central/lib/galaxy/jobs/__init__.py", line 437, in change_state self.sa_session.flush() Regards shashi

1 0

dataset_*_file directory being compied to file_path when running fastqc
by Chorny, Ilya 29 Jul '11

29 Jul '11

I am testing the latest galaxy central distribution by running FASTQC. It creates a directory call dataset_*_file in the job_working_directory and then copies that directory to file_path when the job is complete. Is that on purpose or is that a bug? Thanks, Ilya Ilya Chorny Ph.D. Bioinformatics - Intern ichorny(a)illumina.com<mailto:ichorny@illumina.com> 858-202-4582

2 2

DRMAA options for SGE
by Ka Ming Nip 29 Jul '11

29 Jul '11

Hi, I am trying to configure the proper memory resource requests for my Galaxy tool. This is what I have under the tool_runners section of universe_wsgi.ini [galaxy:tool_runners] ... mytoolname = drmaa://-l mem_free=1G -l mem_token=1G -l h_vmem=1G/ ... When I execute my tool on Galaxy, I get the error below in the shell that I ran "sh run.sh": galaxy.jobs.runners.drmaa ERROR 2011-07-26 09:54:01,930 Uncaught exception queueing job Traceback (most recent call last): File "/home/kmnip/galaxy/galaxy-central/lib/galaxy/jobs/runners/drmaa.py", line 112, in run_next self.queue_job( obj ) File "/home/kmnip/galaxy/galaxy-central/lib/galaxy/jobs/runners/drmaa.py", line 177, in queue_job job_id = self.ds.runJob(jt) File "/home/kmnip/galaxy/galaxy-central/eggs/drmaa-0.4b3-py2.4.egg/drmaa/__init__.py", line 331, in runJob File "/home/kmnip/galaxy/galaxy-central/eggs/drmaa-0.4b3-py2.4.egg/drmaa/helpers.py", line 213, in c File "/home/kmnip/galaxy/galaxy-central/eggs/drmaa-0.4b3-py2.4.egg/drmaa/errors.py", line 90, in error_check DeniedByDrmException: code 17: error: no suitable queues All the flags I used work with qsub commands on the SGE cluster I use. The tool runs when I comment out the line in universe_wsgi.ini. Thanks, Ka Ming

2 4

Please sign me off of the galaxy list
by Mai Huynh 29 Jul '11

29 Jul '11

Dear Galaxy, Please sign me off of the galaxy email list. Even though I have been unsubscribed, but I still get all the galaxy emails. I have to delete them every single day. Thank you very much for your times. Best regards, Mai H.

2 1

organizing histories in galaxy
by Chaolin Zhang 29 Jul '11

29 Jul '11

Hi, Is there a way to organize related histories together into a "project" or folder? I realize in a lot of cases, a pipeline is designed for processing of a single sample, while a study typically consists of multiple samples that go through the same processing. They are then put together for more downstream analysis. It seems to me that it would be logical to put the processing of each individual file into a history, and the combined analysis into another, which would be bundled together. I, and other users here have a growing list of histories, and it is becoming more and more difficult to organize them ... Thanks! Chaolin

2 1