Bug 1317653
Summary: | EINVAL errors while aggregating the directory size by quotad | ||||||
---|---|---|---|---|---|---|---|
Product: | [Red Hat Storage] Red Hat Gluster Storage | Reporter: | Raghavendra Bhat <rabhat> | ||||
Component: | quota | Assignee: | Raghavendra G <rgowdapp> | ||||
Status: | CLOSED ERRATA | QA Contact: | Anil Shah <ashah> | ||||
Severity: | urgent | Docs Contact: | |||||
Priority: | unspecified | ||||||
Version: | rhgs-3.1 | CC: | amukherj, olim, rabhat, rcyriac, rhinduja, rhs-bugs, smohan, vbellur | ||||
Target Milestone: | --- | ||||||
Target Release: | RHGS 3.2.0 | ||||||
Hardware: | x86_64 | ||||||
OS: | Linux | ||||||
Whiteboard: | |||||||
Fixed In Version: | glusterfs-3.8.4-1 | Doc Type: | Bug Fix | ||||
Doc Text: | Story Points: | --- | |||||
Clone Of: | |||||||
: | 1332073 (view as bug list) | Environment: | |||||
Last Closed: | 2017-03-23 05:27:33 UTC | Type: | Bug | ||||
Regression: | --- | Mount Type: | --- | ||||
Documentation: | --- | CRM: | |||||
Verified Versions: | Category: | --- | |||||
oVirt Team: | --- | RHEL 7.3 requirements from Atomic Host: | |||||
Cloudforms Team: | --- | Target Upstream Version: | |||||
Embargoed: | |||||||
Bug Depends On: | |||||||
Bug Blocks: | 1332073, 1351515, 1366746 | ||||||
Attachments: |
|
Comment 2
Vijaikumar Mallikarjuna
2016-03-22 06:49:30 UTC
We found that quotad was in a bad state with memory corruption. After restarting quotad EINVAL was not seen Hi Raghavendra, Have we taken a core-dump and a state-dump, of a quotad process? If taken could you please point to the location of these dump files. We will check the core-dump and see if we can find any clue on why quotad had memory corruption. Thanks, Vijay Created attachment 1174291 [details]
A working quota test script
Hello, I have attached the quota test script that works. The customer would like an engineering review of the script. Thanks & Regards Oonkwee Emerging Technologies RedHat Global Support Hello Manikandan, Patrick Ladd? The script is already in customer hand. This was corrected in a remote session with them. Thanks Hi Oonkwee, Okay then. That's what I heard during the call or at least he was reviewing the scripts. Thanks. Defensive fixes patch is posted upstream: http://review.gluster.org/#/c/14102/ Upstream mainline : http://review.gluster.org/14102 Upstream 3.8 : http://review.gluster.org/15325 And the fix is available in rhgs-3.2.0 as part of rebase to GlusterFS 3.8.4. Test 1a, * Have a dist-rep volume mounted and start I/O on mount point and enable quota. * While I/O is going on, kill one brick of each replica pair. * Bring the brick back up. Wait till self heal completes to the bricks that were down * Once self heal is done, go back to point two above and kill a different brick on each replica pair. * Do this multiple times. * Verify the sizes of directories as shown by Quota both on individual bricks and through quota list command. See whether there are any dependencies Test 1b, * Have a rename heavy workload, just renames or renames of files to which writes are happening. See quota size is accounted fine. As discussed with developers, performed above test. Not seeing any issue quota accounting. But verified on build glusterfs-3.8.4-13.el7rhgs.x86_64 Since the problem described in this bug report should be resolved in a recent advisory, it has been closed with a resolution of ERRATA. For information on the advisory, and where to find the updated files, follow the link below. If the solution does not work for you, open a new bug report. https://rhn.redhat.com/errata/RHSA-2017-0486.html |