Skip to content

yadage: HTCondor and Slurm examples

Tibor Simko requested to merge github/fork/tiborsimko/yad-htc into master

  • HTCondor Yadage example fully working.

  • Slurm Serial example fully working.

  • Slurm Yadage example is having an issue:

$ reana-client status -w yad-hpc
NAME      RUN_NUMBER   CREATED               STARTED               ENDED                 STATUS     PROGRESS
yad-hpc   3            2021-02-17T16:29:52   2021-02-17T16:29:53   2021-02-17T16:32:20   finished   2/2    
$ reana-client logs -w yad-hpc
==> Job logs
==> Step: gendata
...
==> Status: finished
==> Logs:
Auks API request failed : krb5 cred : unable to read credential cache
INFO:    Using cached SIF image
$ reana-client ls -w yad-hpc | grep reana_job
reana_job.945327.out                         0        2021-02-17T16:31:59
reana_job.945327.err                         102      2021-02-17T16:31:59
reana_job.945326.out                         0        2021-02-17T16:31:59
reana_job.945326.err                         102      2021-02-17T16:31:59

That is, the workflow is reported as finished, however there is no stdout, and stderr contains a message about krb5_cred troubles. No data.root or plot.png are generated.

Since Serial Slurm example works well, this may be something particular to the r-w-e-yadage slurm integration? (FWIW, the Serial Slurm example also generates those krb5_cred output messages is stderr.)

Merge request reports