Bug 1266015
Summary: | Not able to recover the corrupted file on Replica volume | ||
---|---|---|---|
Product: | [Red Hat Storage] Red Hat Gluster Storage | Reporter: | Vivek Agarwal <vagarwal> |
Component: | bitrot | Assignee: | Venky Shankar <vshankar> |
Status: | CLOSED DUPLICATE | QA Contact: | Sweta Anandpara <sanandpa> |
Severity: | high | Docs Contact: | |
Priority: | unspecified | ||
Version: | rhgs-3.1 | CC: | annair, asriram, atumball, mzywusko, rabhat, rhs-bugs, sankarshan, smohan, vshankar |
Target Milestone: | --- | Keywords: | ZStream |
Target Release: | --- | ||
Hardware: | Unspecified | ||
OS: | Unspecified | ||
Whiteboard: | |||
Fixed In Version: | Doc Type: | Known Issue | |
Doc Text: |
When an inode is unlinked from the backend (bricks) directly, the corresponding in-memory inode is not cleaned on subsequent lookup. This causes the recovery procedure using healing damons (such as AFR/EC self-heal) to not function as expected as the in-memory inode structure represents a corrupted backend object.
Workaround: The object can be deleted directly from the backend and hence the next lookup on the object fails and the inode is forgotten. Now, the self heal daemon heals the object from the good copy.
|
Story Points: | --- |
Clone Of: | 1238171 | Environment: | |
Last Closed: | 2018-03-13 09:59:15 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: | 1238171 | ||
Bug Blocks: | 1216951, 1238188, 1255604, 1255689, 1266014 |
Comment 3
Anjana Suparna Sriram
2015-09-28 06:47:34 UTC
If the object is deleted directly from the backend, then one does not have to wait till the inode to be forgotten. The next lookup on the object fails and the inode is forgotten right away and lets self heal daemon heal the object from the good copy. Edited further as per our discussion; please review and sign off. Looks good to me. *** This bug has been marked as a duplicate of bug 1238171 *** |