| Summary: | GlusterFS Replicate Feature Doesn't Behave Well On Error | ||
|---|---|---|---|
| Product: | [Community] GlusterFS | Reporter: | Idan Shinberg <idan> |
| Component: | replicate | Assignee: | Pranith Kumar K <pkarampu> |
| Status: | CLOSED DUPLICATE | QA Contact: | |
| Severity: | medium | Docs Contact: | |
| Priority: | low | ||
| Version: | 3.1.1 | CC: | gluster-bugs, vijay |
| Target Milestone: | --- | ||
| Target Release: | --- | ||
| Hardware: | x86_64 | ||
| OS: | Linux | ||
| Whiteboard: | |||
| Fixed In Version: | Doc Type: | Bug Fix | |
| Doc Text: | Story Points: | --- | |
| Clone Of: | Environment: | ||
| Last Closed: | Type: | --- | |
| Regression: | --- | Mount Type: | --- |
| Documentation: | --- | CRM: | |
| Verified Versions: | Category: | --- | |
| oVirt Team: | --- | RHEL 7.3 requirements from Atomic Host: | |
| Cloudforms Team: | --- | Target Upstream Version: | |
|
Description
Idan Shinberg
2010-12-20 14:04:33 UTC
(In reply to comment #0) > In order to test Replicate behavior , I deleted Replicated Files On one of the > Back-ends. As the Result - mount Point access (ls , find or whatever) shows > Nothing. unmounted and mounted again. This time , saw all files but they were > at size zero. Only Issuing find (stating with ls didn't help , as opposed to > older versions ) or actually accessing the files triggered the self heal - but > I guess that's they way it was meant to be( according to gluster documentation > ). > > Still , The No-files-on deletion of one of the back-ends is a bug and it > shouldn't be that way. > > Step to reproduce : > - Create a Replicate Volume using 2 Storage Servers.Back-end File system was > ext3. - Start the volume and remote mount it. > - Create Some Files with contents .( Large and small ). > - Delete File on one of the back-ends. > > Great Tnks. > > Idan hi Idan, We do not support ops done directly on the bricks. Files created directly on the bricks do not have the gfid set on them, and hence all ops on them may fail. If files are deleted from the brick directly, the correct way to recover the files using self-heal is to remount, and perform "find . -noleaf -exec stat '{}' \;" on the mount point. Similar problem is encountered in 2187. so duping. Pranith *** This bug has been marked as a duplicate of bug 2187 *** |