Manually deleting a job? (to free up the queue)

Ask the Community
1

We run a selfhosted QFieldCloud install- I have a delta_apply job which has failed and re-running it is doing nothing. All jobs after this job are also stuck in pending and are now piling up

Is there way to manually delete/flush etc this job?

Output pre

::<<<::08d1334b-8bba-4873-ab80-e87c0bf90c25 Start QGIS Application
07:38:41.664 root      INFO     Starting QGIS app version 33801 (3d4177afc6b)...
07:38:41.977 root      INFO     QGIS app started!
::>>>::08d1334b-8bba-4873-ab80-e87c0bf90c25 2
::<<<::5135a09f-989c-4415-8cfc-512b0b196330 Download Project Directory
07:38:41.978 root      INFO     Preparing a temporary directory for project files…

Timeout error! The job failed to finish within 2000 seconds!

Feedback pre

{
  "container_exit_code": -1,
  "error": "Worker timeout error.",
  "error_class": "",
  "error_origin": "container",
  "error_stack": "",
  "error_type": "TIMEOUT"
}
2

Over 30 minutes is indeed quite high!

Answering your lounge chat question:

Could a QGIS version mismatch cause this? (i.e. re-publishing a project with a higher version number vs that which QFieldCloud is using?).

Possible. Can you clone the instance, update the QGIS version on the clone, and try it to rule it out?

We once had a similar issue caused by a symbology icon loaded from a URL that QFieldCloud couldn’t load for some reason. So it could be anything, really.
If you switch to running with multiple workers (so one stuck project doesn’t get the entire instance stuck), you can make a copy of the project, remove one layer at a time, and let QFieldCloud process it once every time. This way, you can figure out which layer (if any) is causing the issue.