|
From: Laure D. <lau...@to...> - 2012-01-24 17:17:34
|
Chris, Thank you for your help. Yes we are running the pipeline over SGE. The SGE job info from 'show group info' does not help to understand what happened (grid id = 0 and there is no workflow event log). See below an example of 'show group info' we get for a group 'failed' : workflow command id: 7225769 state: failed start time: Tue Jan 17 11:17:52 2012 end time: Tue Jan 17 11:17:53 2012 duration: 1 sec grid id: 0 workflow grid id: 7239862 workflow event log: ? <http://ng6.toulouse.inra.fr/ergatis/cgi/view_formatted_log_source.cgi?file=?> remote wf stderr/sdout :/mnt/wf-working/RunWorkflow.o0,/mnt/wf-working/RunWorkflow.e0 prolog/epilog stderr/stdout /home/guest/staging.*,home/guest/harvesting.* xml: /work/ng6/ergatis/workflow/runtime/bwa_contamination_search/67_default/i1/g137/g137.xml.gz <http://ng6.toulouse.inra.fr/ergatis/cgi/view_formatted_xml_source.cgi?file=/work/ng6/ergatis/workflow/runtime/bwa_contamination_search/67_default/i1/g137/g137.xml.gz> In the working directory, for the job, we have one subdirectory per group of the component. Contents of the subdirectories are described below : - Contents of subdirectories corresponding to a group 'complete' : total 64 -rw-rw-rw- 1 ng6 NG6 530 Jan 17 11:22 event.log -rw-r--r-- 1 ng6 NG6 5 Jan 17 11:19 pid.log -rwxr-xr-x 1 ng6 NG6 7577 Jan 17 11:17 sge_job.sh -rw-r--r-- 1 ng6 NG6 52 Jan 17 11:18 sge_submit.out - Contents of subdirectories corresponding to a group 'failed' : total 16 -rw-r--r-- 1 ng6 NG6 0 Jan 17 11:17 event.log -rw-r--r-- 1 ng6 NG6 7577 Jan 17 11:17 sge_job.sh We also observed that all subdirectories corresponding to a group 'complete' have the permissions 'drwxrwxrwx', whereas those corresponding to a group 'failed' have permissions 'drwxr-xr-x'. Kind regards, Laure Le 20/01/2012 18:48, Chris Hemmerich a écrit : > > Laure, > > Are you running the pipeline over SGE? If so you might be able to get > SGE job info from 'show group info' and look to see if something is > failing in SGE. > > If the directory set in 'workflow_run_dir' from ergatis.ini didn't > provide any clues, you can also check the working directory for the job: > > 'CWD' from workflow/server-conf/sge_mockserver.conf > > Cheers, > Chris > > On Fri, 20 Jan 2012, Laure Devlamynck wrote: > >> Joshua, >> >> Thank you for your help. >> We have found the output and error files. Unfortunately these files >> do not help us to understand why the pipeline have failed. >> >> kind regards, >> Laure >> >> >> Le 18/01/2012 16:50, Joshua Orvis a écrit : >>> Laure - >>> >>> There are a few places to look here. First, there is a log file in >>> the >>> same directory of your pipeline.xml called pipeline.xml.log and >>> another >>> one catching all the standard output called pipeline.xml.run.out. >>> >>> Is this pipeline running jobs on a grid? If so, each of those >>> 'groups' >>> listed corresponds to a job scheduled on your grid, and there are >>> output >>> files for each of them in case your job failed. The path to these is >>> defined by the workflow_run_dir setting you have in your >>> ergatis.ini. For >>> me, that's "/usr/local/scratch/workflow". You can go to that >>> directory >>> and you'll find one folder per pipeline ID. Within that you'll >>> find all >>> the output and error files for each job submission, including those >>> listed >>> in the graphic above. >>> >>> Joshua >>> >>> >>> >>> 2012/1/17 Laure Devlamynck <lau...@to... >>> <mailto:lau...@to...>> >>> >>> Dear all, >>> >>> An error has occured between two command lines within one of our >>> pipeline components, cf. attached file. We have no error message. >>> Moreover this error does not occur each time we submit our >>> pipeline. >>> Does it exist a log file for such an error in order to understand >>> this kind of error ? >>> >>> Thanks in advance for your answers. >>> >>> Kind regards, >>> >>> >>> ------------------------------------------------------------------------------ >>> Keep Your Developer Skills Current with LearnDevNow! >>> The most comprehensive online learning library for Microsoft >>> developers >>> is just $99.99! Visual Studio, SharePoint, SQL - plus HTML5, CSS3, >>> MVC3, >>> Metro Style Apps, more. Free future releases when you subscribe >>> now! >>> http://p.sf.net/sfu/learndevnow-d2d >>> _______________________________________________ >>> Ergatis-users mailing list >>> Erg...@li... >>> <mailto:Erg...@li...> >>> https://lists.sourceforge.net/lists/listinfo/ergatis-users >>> >>> >> >> > > > ------------------------------------------------------------------------------ > Keep Your Developer Skills Current with LearnDevNow! > The most comprehensive online learning library for Microsoft developers > is just $99.99! Visual Studio, SharePoint, SQL - plus HTML5, CSS3, MVC3, > Metro Style Apps, more. Free future releases when you subscribe now! > http://p.sf.net/sfu/learndevnow-d2d > > > _______________________________________________ > Ergatis-users mailing list > Erg...@li... > https://lists.sourceforge.net/lists/listinfo/ergatis-users |