[
http://jira.jboss.com/jira/browse/JBMESSAGING-1116?page=all ]
Ovidiu Feodorov updated JBMESSAGING-1116:
-----------------------------------------
Summary: Consuming client grinds to a halt if a node different from the
"active" one fails (was: Fail-over does not complete if a node different from
the "active" one fails)
Consuming client grinds to a halt if a node different from the
"active" one fails
---------------------------------------------------------------------------------
Key: JBMESSAGING-1116
URL:
http://jira.jboss.com/jira/browse/JBMESSAGING-1116
Project: JBoss Messaging
Issue Type: Bug
Components: JMS Clustering
Affects Versions: 1.4.0.GA
Reporter: Ovidiu Feodorov
Assigned To: Tim Fox
Priority: Critical
Fix For: 1.4.0.SP1
This is how to replicate it:
- Load a queue with a large number of messages (I used 100,000 10K text persistent
messages), by connecting to node A
- Start consuming concurrently using 50 parallel consumers over a connection to node A
- Kill node B
The consumption rate drop to 0 and never recovers.
Trying to close the connection on the client-side never completes.
Trying to shut down node A fails, need to eventually kill it.
--
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators:
http://jira.jboss.com/jira/secure/Administrators.jspa
-
For more information on JIRA, see:
http://www.atlassian.com/software/jira