Project

General

Profile

Actions

Bug #15513

closed

RGW multisite: Delete requests do not propagate to the peer zone

Added by Shilpa MJ about 8 years ago. Updated about 7 years ago.

Status:
Resolved
Priority:
Normal
Assignee:
-
Target version:
-
% Done:

0%

Source:
other
Tags:
Backport:
Regression:
No
Severity:
2 - major
Reviewed:
Affected Versions:
ceph-qa-suite:
Pull request ID:
Crash signature (v1):
Crash signature (v2):

Description

In an active-active multisite config, after syncing a file to the destination node, a delete on the file from the source or destination node does not sync to the peer node.

1. Configured multisite active-active sites with rgw1 as master gateway node
2. Sync a file from rgw1 node to rgw2 node
3. Now once the sync is complete, delete the file from either rgw1 or rgw2 node
4. The delete operation is not synced to the peer zone

On the peer zone where the delete operation sync failed (rgw2):

2016-04-15 09:33:03.130373 7f82b9ffb700 0 _send_request(): deleting obj=container4:ks-post-nochroot.log
2016-04-15 09:33:03.133463 7f82b9ffb700 20 _send_request(): skipping object removal obj=container4:ks-post-nochroot.log (obj mtime=2016-04-15 09:32:12.0.081866s, request timestamp=2016-04-15 09:32:05.0.101811s)

realm 4e00a610-36e9-43d0-803e-4001442b8232 (earth)
zonegroup e66e1293-e63b-4afe-9dad-3397647dfb03 (us)
zone 001da65b-c3a8-42e2-a1ce-79cacefbace2 (us-2)
metadata sync syncing
full sync: 0/64 shards
metadata is caught up with master
incremental sync: 64/64 shards
data sync source: acadcc66-10b9-4829-b8e2-306c0048bff5 (us-1)
syncing
full sync: 0/128 shards
incremental sync: 128/128 shards
data is behind on 1 shards
oldest incremental change not applied: 2016-04-15 10:06:35.0.959384s

Related issues 1 (0 open1 closed)

Related to rgw - Bug #16464: multisite sync races with deletesResolvedCasey Bodley06/23/2016

Actions
Actions #1

Updated by Robin Johnson about 8 years ago

  • Project changed from Ceph to rgw
Actions #2

Updated by Abhishek Lekshmanan over 7 years ago

almost all fixes on obj/bucket deletes should have already been backported to jewel by now

Actions #3

Updated by Casey Bodley about 7 years ago

  • Related to Bug #16464: multisite sync races with deletes added
Actions #4

Updated by Casey Bodley about 7 years ago

  • Status changed from New to Resolved
Actions

Also available in: Atom PDF