Re: [galaxy-dev] Galaxy with PBS runner Setting

2012-01-03 Thread Nate Coraor
On Dec 20, 2011, at 12:19 AM, Le-Shin Wu wrote:

 Thanks a million, Nate. I was able to submit my PBS job though galaxy. But I 
 got another problem once the PBS job is finished. 
 
 I submit a bowtie alignment testing job to our cluster by using PBS runner 
 through galaxy. But I always got an error message showing on the right 
 history panel as below.
 
 An error occurred running this job:
 # reads processed: 17214026
 # reads with at least one reported alignment: 16405756 (95.30%)
 # reads that failed to align: 808270 (4.70%)
 Reported 16405756 alignments to 1 output stream(s)
 
 In fact, the message showing here is the output message from bowtie 
 application and I did get the correct alignment result stored on galaxy 
 server machine. Therefore, I tried adding some PBS options, such as -k eo 
 or -j eo -o /dev/null, to the PBS runner in universe_wsgi.ini to remove the 
 standard output message from bowtie. But this time I got the following error 
 message. 

Hi Le-Shin,

I am not familiar enough with bowtie to know why this error occurred.  Perhaps 
someone else on the list can reply with some information.  Please keep replies 
on the mailing list to ensure this is possible.

 An error occurred running this job:Job output not returned by PBS: the output 
 datasets were deleted while the job was running, the job was manually 
 dequeued or there was a cluster error.  
 
 It looks like if galaxy PBS runner did not receive the standard output from 
 cluster, it will throw an error. Could you please help us to solve this 
 problem. Thanks a lot.

That's correct, Galaxy controls the location of the stdout/stderr files, 
changing these will cause such failures.

--nate

 
 Le-Shin 
 
 
 


___
Please keep all replies on the list by using reply all
in your mail client.  To manage your subscriptions to this
and other Galaxy lists, please use the interface at:

  http://lists.bx.psu.edu/


Re: [galaxy-dev] Galaxy with PBS runner Setting

2011-12-19 Thread Nate Coraor
On Dec 19, 2011, at 3:03 PM, Le-Shin Wu wrote:

 We are trying to get galaxy to work on our cluster. Unfortunately, there are 
 some questions I still cannot find the answer to and I'm hoping someone here 
 can help me out. 
 
 We would like to use the unified approach with TPRQUE PBS to support running 
 galaxy tools on a cluster. Our questions are listed as below.
 
 (1) How to configure the universe_wsgi.ini file? Based on the galaxy wiki 
 page, we changed the following parameters in universe_wsgi.ini
  new_file_path =  /N/dc/projects/galaxy/new_files # this directory is 
 accessible both to galaxy server and cluster nodes
  start_job_runners = 'pbs'
  default_cluster_job_runner =  pbs://cluster.hostname.edu/ 
 #cluster.hostname.edu is the hostname for our cluster
 
 Also under [galaxy:tool_runners] section, we add a new tool named mytool 
 for testing as below.
 mytool = pbs://cluster.hostname.edu/ 
 
 Does this configuration look correct and enough?
 
 (2) We created a new python egg for pbs by using
  LIBTORQUE_DIR=/usr/local/lib/libtorque python scripts/scramble.py -e 
 pbs_python --config=universe_wsgi.ini
 
  A new file pbs_python-4.1.0-py2.6-linux-x86_64-ucs4.egg was then created 
 under eggs. 
 
  Do we need to do other things or only this operation is enough?
 
 (3) We created a new directory named mytool under  tools and added a new file 
 named newTool.xml in it. The content of newTool.xml is as below. Note that, 
 the /N/soft/mason/test_tool is a executable binary file at our cluster. 
 
 tool id=mytool name=Testing tool version=0.12.7
  descriptiondo the pbs testing/description
  command
  /N/soft/mason/test_tool -p 8 hg18 -f $inputseq $output
  /command
  inputs
 param name=db type=inputseq  label=input file /
 /inputs
  outputs
data format=sam name=output label=result /
  /outputs
 tests
 test
  param name=inputseq value=a.fa  /
  output name=output file=b.sam  /
 
/test
  /tests
 
  help
  **What it does**
   This tool uses Bowtie to produce an alignment.
 .. _Bowtie: http:
 
  /help
 /tool
 
 Does this xml file look correct? Do we also need a python wrapper file to 
 somehow invoking PBS support?
 
 (4) Add the following text into tool_conf.xml.
 section name=My Tools id=mt
  tool file=mytool/newTool.xml /
 /section
 
 (5) After doing the above modification and restarting our galaxy server , we 
 can see My Tools showing on the tools panel of our galaxy web interface. 
 But after we select an input file and click execute button, the job status 
 shows an error Unable to run job due to a misconfiguration of the Galaxy job 
 running system.  Please contact a site administrator. We also check the log 
 file, it shows  Invalid job runner: bps.

Hi Le-Shin,

It would appear that the PBS job runner failed to start.  You should be able to 
see a traceback with more details about this displayed during the server's 
startup.

--nate

 
 We appreciate very much for your kind help.
 
 
 Le-Shin  
 ___
 Please keep all replies on the list by using reply all
 in your mail client.  To manage your subscriptions to this
 and other Galaxy lists, please use the interface at:
 
  http://lists.bx.psu.edu/


___
Please keep all replies on the list by using reply all
in your mail client.  To manage your subscriptions to this
and other Galaxy lists, please use the interface at:

  http://lists.bx.psu.edu/