[Pegasus-users] pegasus 4.0 errors in running pegasus 3.1 workflows

[Yu Huang]

That was a stupid error on my part. I installed pegasus on two machines but
not the 3rd one.

It's now working.

thanks,
yu


On Wed, Mar 14, 2012 at 3:53 AM, Karan Vahi <vahi at isi.edu> wrote:

> Hi Yu
>
> It looks like on the cluster nodes, pegasus worker package is not installed
>
> I gather you installed 4.0 using rpm on the submit node
> /usr/bin/pegasus-kickstart
>
> /usr/bin is a local path, that most probably in your setup is only
> accessible on the submit node.
>
> You did not install it on the worker nodes. Can you install it ?
>
> Alternatively, you can get pegasus to stage the worker package for you.
> http://pegasus.isi.edu/wms/docs/4.0/reference.php#id748665
>
> Karan
>
>
>
>
>
> On Mar 14, 2012, at 12:45 AM, Yu Huang wrote:
>
> Hi guys,
>
> As far as i understand from the 4.0 announcement, if i'm running on shared
> filesystem, old workflow xmls should still work. but i ran into 2 problems
> after having submitted the workflows.
>
> 1. Job below is folder creation job as I want to put some output files in
> specific folders. "/home/crocea/script/pymodule/shell/mkdirWrap.sh" is
> shell wrapper around mkdirhier. somehow pegasus complained "No such file or
> directory" while I could just run it in shell. In another workflow i ran,
> this held job eventually went through (same error initially). i have no
> idea of what's going on.
>
> STAT  IN_STATE  JOB
> Run      11:50  GW_LD_pattern_between_call32_call80Type3Peaks_EMMAX-0
> Held     02:36   ┗━mkdirWrap_ID0000001
>  ┗━Error from slot1 at banyan: Failed to execute
> '/usr/bin/pegasus-kickstart' with arguments -n workflow::mkdirWrap:1.0 -N
> ID0000001 -R condorpool -B 4000000 -L
> GW_LD_pattern_between_call32_call80Type3Peaks_EMMAX -T
> 2012-03-14T00:20:16-07:00 /home/crocea/script/pymodule/shell/mkdirWrap.sh
> GWLD: (errno=2: 'No such file or directory')
> Summary: 2 Condor jobs total (R:1 H:1)
>
> UNREADY   READY     PRE  QUEUED    POST SUCCESS FAILURE %DONE
>   1,734       0       0       2       0       2       0   0.1
> Summary: 1 DAG total (Running:1)
>
>
> 2. this problem is related to job clustering.
>
> STAT  IN_STATE  JOB
>
> Run      11:57  GW_LD_pattern_between_call32_call80Type1Peaks_KW-0
>
> Held     02:17
> ┣━merge_workflow-OutputLociIDOfResultPeakInHDF5.py-1.0_PID3_ID52
>  ┗━Error from slot1 at banyan: Failed to execute '/usr/bin/pegasus-cluster'
> with arguments -f: (errno=2: 'No such file or direct
> ory')
> Held     02:16
> ┣━merge_workflow-OutputLociIDOfResultPeakInHDF5.py-1.0_PID3_ID50
>  ┗━Error from slot1 at banyan: Failed to execute '/usr/bin/pegasus-cluster'
> with arguments -f: (errno=2: 'No such file or direct
> ory')
> ...
>
> I attached the cluster job's "sub" and "in" file. let me know if you need
> more files.
>
> thanks,
> yu
>
> --
> http://www-scf.usc.edu/~yuhuang
>  <merge_workflow-OutputLociIDOfResultPeakInHDF5.py-1.0_PID3_ID52.sub><
> merge_workflow-OutputLociIDOfResultPeakInHDF5.py-1.0_PID3_ID52.in>
> _______________________________________________
> pegasus-users mailing list
> pegasus-users at isi.edu
> http://mailman.isi.edu/mailman/listinfo/pegasus-users
>
>
>


-- 
http://www-scf.usc.edu/~yuhuang
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mailman.isi.edu/pipermail/pegasus-users/attachments/20120314/4a909249/attachment.html