Install_ancil failing

1

HI,

I have two very related jobs which are both failing in install_ancil. They both are using prebuild which points to the same directory. When they run the workflow instal_ancil is failing and on puma/cylc gui I can’t see any .err or .out files – see /home/n02/n02/tetts/cylc-run/u-ds995/run2/log/job/20100901T0000Z/install_ancil/02 (or 01)

On archer2 I can see the log files which tells me that run_ancil is running out of time. See /work/n02/n02/tetts/cylc-run/u-ds995/run2/log/job/20100901T0000Z/install_ancil/02/job.err (or 01).

(ModelOptimisation) tetts@ln01:/work/n02/shared/tetts/test_run> more /work/n02/n02/tetts/cylc-run/u-ds995/run2/log/job/20100901T0000Z/install_ancil/02/job.err
slurmstepd: error: *** JOB 11015280 ON dvn01 CANCELLED AT 2025-10-02T10:29:57 DUE
TO TIME LIMIT ***

Everyone should have read access to those files.

So, what should I do? I suspect there is some temp file system problem somewhere…

Simon

2

I tried again – releasing the instal_ancil task and seems to have worked. I guess a glitch somewhere in the systems….

Simon

3

try giving it more time, PT2M maybe?

[[INSTALL_RESOURCE]]
    inherit = HPC_SERIAL
    execution time limit = PT1M
4

I had similar problems with other cases and reported to archer2 help desk. They told me it was because the n02 quota was full. And they’d tell n02 so!

Simon