[Ncep.hmon] Alternative Rocoto to work around "unavailable" jobs

Samuel Trahan - NOAA Affiliate samuel.trahan at noaa.gov
Mon Apr 29 21:15:07 UTC 2019


Hi,

There are two bugs preventing Rocoto from finding jobs.

1) The sbatch command sometimes reports a job was not submitted when it
really was submitted.  The admins will increase a timeout to reduce this
problem.

2) The squeue -j command has a 64 character limit, which is violated for
Rocoto workflows that have more than about ten jobs queued or running at a
time.  You need a patched Rocoto to work around this.

The central Rocoto 1.3.0-RC5 will be patched as soon as I can find someone
with root access on Jet and Theia.  Until then:

    JET: module use /lfs3/projects/hwrf-vd/soft/modulefiles
    THEIA: module
use /scratch4/NCEPDEV/nems/noscrub/emc.nemspara/soft/modulefiles

    module load rocoto/1.3.0-RC5-smallj

Sincerely,
Sam Trahan
-------------- next part --------------
An HTML attachment was scrubbed...
URL: https://www.lstsrv.ncep.noaa.gov/pipermail/ncep.hmon/attachments/20190429/75618bff/attachment.html 


More information about the Ncep.hmon mailing list