Bug 874298 - HA add UUID tag to avoid using an out of date queue.
HA add UUID tag to avoid using an out of date queue.
Product: Red Hat Enterprise MRG
Classification: Red Hat
Component: qpid-cpp (Show other bugs)
Unspecified Unspecified
high Severity high
: 2.3
: ---
Assigned To: Alan Conway
MRG Quality Engineering
: OtherQA
Depends On:
Blocks: 698367
  Show dependency treegraph
Reported: 2012-11-07 16:09 EST by Alan Conway
Modified: 2013-03-19 12:38 EDT (History)
5 users (show)

See Also:
Fixed In Version: qpid-cpp-0.18-9
Doc Type: Bug Fix
Doc Text:
Story Points: ---
Clone Of:
Last Closed: 2013-03-19 12:38:20 EDT
Type: Bug
Regression: ---
Mount Type: ---
Documentation: ---
Verified Versions:
Category: ---
oVirt Team: ---
RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: ---

Attachments (Terms of Use)

External Trackers
Tracker ID Priority Status Summary Last Updated
Apache JIRA QPID-4428 None None None 2012-11-07 16:13:05 EST

  None (edit)
Description Alan Conway 2012-11-07 16:09:56 EST
Description of problem:

Here's the scenario: we have a cluster with primary A and backups B and C. There is a queue Q that is replicated to both. Now A dies and B takes over as primary. Before C has connected to B, a client destroys Q and creates a new queue with the same name, Q. When C does connect it sees a queue named Q and mistakenly assumes it is the same queue as it's own Q - so it has an incorrect replica of Q. The same scenario applies to exchanges.

Version-Release number of selected component (if applicable):

How reproducible: 100%

Steps to Reproduce:
1. Start a 3 cluster
2. Create a queue Q with messages
3. Kill the primary but don't promote any backups
4. On one broker delete and re-create Q, then promote the broker
Actual results: The other backup will keep it's out-of-date Q and messages.

Expected results: The other backup deletes its old Q and creates a new one to replicate the new Q.

Additional info:

The unit test in this patch illustrates how to reproduce the problem.

Note You need to log in before you can comment on or make changes to this bug.