Note: This bug is displayed in read-only format because the product is no longer active in Red Hat Bugzilla.
This project is now read‑only. Starting Monday, February 2, please use https://ibm-ceph.atlassian.net/ for all bug tracking management.

Bug 1225222

Summary: [Release Note] OSD crash in release_op_ctx_locks with rgw and pool snaps
Product: [Red Hat Storage] Red Hat Ceph Storage Reporter: Vasu Kulkarni <vakulkar>
Component: DocumentationAssignee: John Wilkins <jowilkin>
Status: CLOSED CURRENTRELEASE QA Contact: ceph-qe-bugs <ceph-qe-bugs>
Severity: high Docs Contact:
Priority: high    
Version: 1.3.0CC: ceph-eng-bugs, ceph-qe-bugs, hnallurv, jowilkin, kdreyer, mlawrenc, ngoswami, sjust, smanjara, vakulkar
Target Milestone: rc   
Target Release: 1.3.0   
Hardware: Unspecified   
OS: Unspecified   
Whiteboard:
Fixed In Version: Doc Type: Bug Fix
Doc Text:
Story Points: ---
Clone Of: 1223941 Environment:
Last Closed: 2015-10-21 14:34:40 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: 1223941    
Bug Blocks: 1230323    

Comment 2 Harish NV Rao 2015-05-28 13:08:32 UTC
Hi Vasu, 
>Release flag is set for 1.3.1, but this should be documented in Release notes for 1.3.0 as Known issue so that customers dont try this config. 

- Please share which config should not be tried and the reason for release note request on this BZ.

Comment 3 Samuel Just 2015-05-28 16:08:54 UTC
The reason is in the above comment.  Users must not use pool snapshots on a pool used by radosgw as it will crash the osd.

Comment 4 Harish NV Rao 2015-05-28 16:41:25 UTC
Thanks Sam! Will this be fixed in code in future release?

Comment 5 Samuel Just 2015-05-28 17:26:18 UTC
Yeah, I have a patch for it, it's not in master yet.  I'm working on improving certain ceph-qa-suite tests to exercise this case.

Comment 6 Ken Dreyer (Red Hat) 2015-05-29 22:01:03 UTC
John, is there a particular document that QE should be looking at in order to review this and move this bz to VERIFIED?

Comment 8 John Wilkins 2015-07-13 22:10:13 UTC
All outstanding bugs were added to the release notes.

https://access.redhat.com/beta/documentation/en/red-hat-ceph-storage-13-release-notes

Comment 9 Harish NV Rao 2015-10-08 10:41:02 UTC
This defect is found under known issues section in the doc link shared by John in comment 8. Marking this as verified from the doc perspective for release 1.3.0. 

But the original issue need to be fixed in later releases (see comment 5). We may have to create a new defect for this. 

Sam, Can you please let me know if this needs a separate ticket?

Comment 10 Samuel Just 2015-10-16 03:18:24 UTC
The fix appears in master, infernalis, and hammer, but not v0.94.2.  I guess that means it needs another ticket (unless it was cherry-picked in for the most recent release)?

The hammer backport was

commit c7b6a6370a69149ea94f9e35d536aa90f06e7659
Author: Samuel Just <sjust>
Date:   Tue May 19 10:56:11 2015 -0700

    ReplicatedPG::finish_ctx: take excl lock if operation is rw
    
    Fixes: #11677
    Signed-off-by: Samuel Just <sjust>
    (cherry picked from commit 5c2b795724423ed484ab451de855ddcfc085342b)

and two other test commits

1550a569dab120ce28396fe365565e8e4acd9801
4cdc5f7d6b3ec488c79c09cb44a43d4d9398b74c

This bug does not appear to be applicable to firefly.

Comment 11 Harish NV Rao 2015-10-16 12:23:04 UTC
Ken,

> The fix appears in master, infernalis, and hammer, but not v0.94.2.  I guess that means it needs another ticket (unless it was cherry-picked in for the most recent release)?

Can you please confirm whether this was picked up for 0.94.3 or not? If not, i will create a separate BZ for it...

Comment 12 Ken Dreyer (Red Hat) 2015-10-19 14:25:30 UTC
(In reply to Harish NV Rao from comment #11)
> Can you please confirm whether this was picked up for 0.94.3 or not? If not,
> i will create a separate BZ for it...

Yes, the backport is in v0.94.3.

Comment 13 Harish NV Rao 2015-10-21 08:31:13 UTC
We will be verifying the fix for this in 1.3.1 as part of Bug #1223941.

Comment 14 Harish NV Rao 2015-10-21 08:39:16 UTC
For 1.3.0, the Doc fix has been already verified.