Bug #11148

[Crunch2] Propagate dispatch error messages (e.g., sbatch fails) to user via logs/websocket

Added by Tom Clegg 9 months ago. Updated 3 months ago.

Status:NewStart date:02/21/2017
Priority:NormalDue date:
Assignee:-% Done:

0%

Category:Crunch
Target version:Arvados Future Sprints
Story points-
Velocity based estimate-

Description

Background

If crunch-dispatch-slurm encounters an error while trying to submit a container to the slurm queue, it logs a message to stderr, but there is no way for a user to see this. If the error persists, the user has no way of knowing whether anything is happening.

Fix

In addition to printing a message on stderr, crunch-dispatch-slurm should use the Arvados API to emit a log that is visible to the user.
  • object_uuid: container uuid
  • event_type: "dispatch"

Similar to source:services/crunch-run/logging.go#L183

A log should also be emitted when the container is successfully submitted to the slurm queue.


Related issues

Related to Arvados - Bug #10704: [Crunch2] sbatch submit failures not reported to use, loo... Resolved 01/31/2017
Related to Arvados - Bug #10729: [Crunch2] Propagate error messages if sbatch command succ... New 12/14/2016

History

#1 Updated by Tom Clegg 9 months ago

  • Description updated (diff)

#2 Updated by Tom Clegg 9 months ago

  • Description updated (diff)

#3 Updated by Tom Morris 3 months ago

  • Target version set to Arvados Future Sprints

Also available in: Atom PDF