Bug 710364 - Restart of corosync causes qpidd shutdown
Summary: Restart of corosync causes qpidd shutdown
Keywords:
Status: CLOSED NOTABUG
Alias: None
Product: Red Hat Enterprise MRG
Classification: Red Hat
Component: qpid-cpp
Version: 2.0
Hardware: x86_64
OS: Linux
medium
high
Target Milestone: 2.2
: ---
Assignee: mick
QA Contact: MRG Quality Engineering
URL:
Whiteboard:
Depends On:
Blocks:
TreeView+ depends on / blocked
 
Reported: 2011-06-03 07:52 UTC by Pavel Moravec
Modified: 2011-11-15 19:46 UTC (History)
1 user (show)

Fixed In Version:
Doc Type: Bug Fix
Doc Text:
Clone Of:
Environment:
Last Closed: 2011-11-15 19:46:28 UTC
Target Upstream Version:


Attachments (Terms of Use)
qpidd debug log from reproduction (7.61 KB, application/x-gzip)
2011-06-03 07:52 UTC, Pavel Moravec
no flags Details

Description Pavel Moravec 2011-06-03 07:52:27 UTC
Created attachment 502747 [details]
qpidd debug log from reproduction

Description of problem:
Having MRG / qpidd in cluster, a simple restart of openais causes automatic qpidd process shutdown.

Version-Release number of selected component (if applicable):
any (checked on 1.3 and on 2.0 devel version)

How reproducible:
100%

Steps to Reproduce:
1. Setup qpidd in cluster
2. service openais start
3. service qpidd start
4. sleep 30    (just to be sure qpidd is fully up)
5. service openais restart
6. service qpidd status
  
Actual results:
Step 6 outputs:
qpidd dead but pid file exists

Expected results:
qpidd is robust enough to overcome openais restart

Additional info:
qpidd debug logs attached, the most relevant part is:
..
2011-06-03 09:49:50 debug Exception constructed: Error in CPG dispatch: library (2)
2011-06-03 09:49:50 critical Error in cluster dispatch: Error in CPG dispatch: library (2)
2011-06-03 09:49:50 notice cluster(10.34.1.222:19533 LEFT) leaving cluster guests
2011-06-03 09:49:50 debug Shutting down CPG
2011-06-03 09:49:50 notice Shut down
..

Comment 2 mick 2011-11-15 19:46:28 UTC
Sorry, it is not feasible to re-engineer current cluster code to seamlessly survive an openais failure/restart.  

It *is* possible that a future version of High-Availability Qpid will not require openais at all, but for versions that use openais we will require a qpidd restart after any openais restart.


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