Opened 2 weeks ago

Closed 6 days ago

#2664 closed help (fixed)

glm_linkprv submit failed

Reported by: amenon Owned by: um_support
Priority: normal Component: UM Model
Keywords: submit failed Cc:
Platform: ARCHER UM Version: 10.9

Description

Hi,

glm_linkprev of one of my suites u-bc643 in Archer is failing to get submitted since yesterday. Archer service status shows that low priority jobs are disabled since yesterday. Could you please let me know if this is the reason for the submit-fail. However, the job activity log gives the following error:

qsub: request rejected as filter hook 'update_user_environment' encountered an exception. Please inform Admin

And Archer troubleshooting guide (http://www.archer.ac.uk/documentation/troubleshooting.php) says that this error is encountered if there is no select statement in the job script. I don't quite understand what that means. Should I wait for the low priority jobs to be enabled again or do I need to do something about this 'select statement' to get over this error? The suite succeeded some other cycles without encountering this error earlier.

Cheers,
Arathy

Change History (8)

comment:1 Changed 2 weeks ago by grenville

Arathy

Has glm_linkprev ever run OK on ARCHER?

It looks like there should be a directive like

#PBS -l select=serial=true:ncpus=1

in the job file, but only

#PBS -q serial

is present. Please check the HOST_HPC section of the suite.rc file.

Grenville

comment:2 Changed 2 weeks ago by amenon

Hi Grenville,

Thanks. Yes, glm_linkprev ran fine in Archer for other cycles in this suite and for several cycles in a copy of this suite. I checked suite.rc file. It only has the job submission delay time in the HOST_HPC section. How can I get around this? Can I add "-l select=1:ncpus=1:mem=1GB" under the directives of HOST_HPC in the suite-adds.rc file?

Cheers,
Arathy

comment:3 Changed 2 weeks ago by grenville

Please point me to a log file for a successful run of glm_linkprev on ARCHER

Grenville

comment:4 Changed 2 weeks ago by amenon

Hi Grenville,

This is the directory where the log files to a succeeded cycle are there:
/work/n02/n02/amenon/cylc-run/u-bc643/log/job/20160701T0000Z

However, I don't find the log files for glm_linkprev in there. DOn't understand why.

Cheers,
Arathy

comment:5 Changed 2 weeks ago by amenon

Hi Grenville,

Here is the path to a log file for a successful run of glm_linkprev in Archer from one of my older suites.

/work/n02/n02/amenon/cylc-run/u-ai540/log/job/20160930T0000Z/glm_linkprev/01

Cheers,
Arathy

comment:6 Changed 2 weeks ago by grenville

If you look in

/home/n02/n02/amenon/cylc-run/u-ai540/suite.rc.processed, you'll see the entry

glm_linkprev?

[directives?]

-l walltime = 00:20:00
-l select=1:ncpus=1:mem=1GB

but /home/n02/n02/amenon/cylc-run/u-bc643/suite.rc.processed

does not have that entry hence failure.

Try adding

{{DRV_MOD["name"]}}_linkprev?

[directives?]

-l walltime = 00:20:00
-l select=1:ncpus=1:mem=1GB

to /home/amenon/roses/u-bc643/site/ncas-cray-xc30/suite-adds.rc

Grenville

comment:7 Changed 2 weeks ago by amenon

Hi Grenville,

Thanks a lot. It got submitted now.

Arathy

comment:8 Changed 6 days ago by willie

  • Resolution set to fixed
  • Status changed from new to closed
Note: See TracTickets for help on using tickets.