Feature #373
closed
mechanism to remove broken cluster jobs from showing in the menu
Added by Anchi Cheng over 14 years ago.
Updated over 14 years ago.
Description
When a cluster job does not start up right, it is left without appionjob.log and will show as running but it is in fact terminated at the cluster. Need some way to remove them from the query
Related issues
1 (1 open — 0 closed)
- Status changed from New to In Code Review
- Assignee set to Amber Herold
1.Added the option button if no job is in cluster but the job is still have running or queued status. revision r13583
2.Make broken queued runs link to the same page and aborting mechanism. revision r13584
Testing:
1. ssh to guppy
2. modify your sinedon.cfg so that the database name is wrong. For example, make [leginondata] db: nonsense
3. from a processing page, login on the top right corner
4. run something simple such as DogPicking on the images of the session. sent the job to guppy to run,
5. check status of the job. It should say Queued but the log would show error, and quit without changing the status in the database. Also the menu would show 1 queued under DogPicking
6. click on "Kill Job" below the log.
7. the page would refresh and indicate that the job is aborted.
8. refresh the page again the menu should not show that any job is queued
Don't forget to change your sinedon.cfg on guppy back.
- % Done changed from 0 to 100
- Status changed from In Code Review to In Test
- Status changed from In Test to Closed
Also available in: Atom
PDF