phillxnet
(Philip Guyton)
August 7, 2017, 7:23pm
8
@Christian_Rost and @Dragon2611 from:
So I had scheduled scrub task from when I first built the VM, sadly at some point it got stuck and I resumed it from the CLI.
The problem is Rockstor thinks it’s still running so doesn’t start a new one, meaning the Pool hasn’t been scrubbed for nearly a year (Opps!)
[root@rock-home ~]# btrfs scrub status /mnt2/Pool1/
scrub status for 48e9fe04-9f49-4c33-b9d7-12aefe745674
scrub resumed at Fri Aug 19 22:10:20 2016 and finished after 90:42:26
total bytes scrubbed: 5.46TiB with 0 errors
423 Au…
As per:
I ended up breaking that issue out on it’s own:
opened 05:44PM - 07 Aug 17 UTC
closed 11:25PM - 08 Aug 17 UTC
bug
Thanks to forum members arbrick, Christian_Rost and Dragon2611 for helping to br… ing this bug to light.
The intention with the force option for scrubs and balances is two fold:
1. change the status of the last listed job (if currently 'started' or 'running') to 'terminated'.
2. pass the -f option to the respective btrfs commands and monitor this under a new job.
It is surmised that a recent Django or related update altered the requirement by which the existing 'force' option was passed from the Web-UI front end to the backend balance/scrub command wrappers. The result was that the force option was ineffectual in both 1. and 2. above. However in the case of a scrub there existed an additional bug that prevented 2. above. This is also fixed in the pr to be associated with this issue. Tasks addressed by the pending pr:
- [x] Fix 'force' UI option transition to backend - scrub/balance
- [x] Fix -f addition to scrub command with force option.
Please update the following forum threads with the resolution of this issue:
https://forum.rockstor.com/t/scrub-fails-to-end-fails-to-restart-after-shutdown/2769
https://forum.rockstor.com/t/scrub-wont-start-unfinished-job-in-history-table/3588
https://forum.rockstor.com/t/scheduled-scrub-not-running/3608
and both these issues now have associated pull requests so upon the review process going as hoped these fixes should be in place in the testing channel updates soon.
Hope that helps.