Bug 786802

Summary: Failover is not done correctly with MDB
Product: Red Hat Enterprise MRG Reporter: Jiri Sedlacek <jsedlace>
Component: qpid-jcaAssignee: Weston M. Price <wprice>
Status: CLOSED CURRENTRELEASE QA Contact: MRG Quality Engineering <mrgqe-bugs>
Severity: high Docs Contact:
Priority: high    
Version: 2.0CC: cdewolf, iboverma, jross, oskutka, tross, wprice
Target Milestone: ---   
Target Release: ---   
Hardware: Unspecified   
OS: Unspecified   
Whiteboard:
Fixed In Version: Doc Type: Bug Fix
Doc Text:
Story Points: ---
Clone Of: Environment:
Last Closed: 2013-03-12 12:07:00 UTC Type: ---
Regression: --- Mount Type: ---
Documentation: --- CRM:
Verified Versions: Category: ---
oVirt Team: --- RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: --- Target Upstream Version:
Bug Depends On: 834416    
Bug Blocks:    

Description Jiri Sedlacek 2012-02-02 12:33:44 UTC
Description of problem:
When MDB is deployed to EAP instance and connected to cluster of MRG nodes, failover is not done correctly. 


Steps to Reproduce:

1/ run cluster of 2 or more mrg nodes
2/ configure jboss-eap-5.1.2 to use xa recovery with mrg (in jbossts-properties.xml), deploy qpid-ra.rar with embeded qpid-xa-recoveryjboss.jar
3/ prepare and deploy message-driven bean with anotations below, use your own mrg server/queue:

@MessageDriven(mappedName = "jms/QpidListener", activationConfig = { @ActivationConfigProperty(propertyName = "acknowledgeMode", propertyValue = "Auto-acknowledge"),
      @ActivationConfigProperty(propertyName = "destinationType", propertyValue = "javax.jms.Queue"),
      @ActivationConfigProperty(propertyName = "destination", propertyValue = "queue/mrg_ha_kill_mrg_in"),
      @ActivationConfigProperty(propertyName = "connectionURL", propertyValue = "amqp://guest:guest@/test?brokerlist='tcp://mrg01.mw.lab.eng.bos.redhat.com:5677?sync_ack='true'&max_prefetch='0''&failover='failover_exchange'"),
      @ActivationConfigProperty(propertyName = "useLocalTx", propertyValue = "false"),
      @ActivationConfigProperty(propertyName = "maxSession", propertyValue = "15") })
@ResourceAdapter("qpid-ra.rar")

4/ start eap instance
5/ send 1000 (or more) messages to mdb queue
6/ kill the mrg instance your mdb is connected to

  
Actual results:

Failover should be done and mdb should be reconnected to another mrg node. Instead of this, in eap instance server.log you can see something like this: 

12:27:59,029 INFO  [FailoverExchangeMethod] ============================================================
12:27:59,029 INFO  [FailoverExchangeMethod] FailoverExchange:

 Current Broker Index:0
 Failed Attempts:0
 Orignal broker details:tcp://mrg01.mw.lab.eng.bos.redhat.com:5677?sasl_mechs='ANONYMOUS'

 -------- Broker List -----------
>tcp://mrg01.mw.lab.eng.bos.redhat.com:5677?sasl_mechs='ANONYMOUS'
--------------------------------

12:27:59,029 INFO  [FailoverExchangeMethod] FailoverAllowed true
12:27:59,029 INFO  [FailoverExchangeMethod] ============================================================

The only node of cluster seems to be the killed node, which is wrong (it's the node, which is configured in MDB)


Expected results:

MDB should receive correct cluster topology, process it and reconnect to one of other available nodes.

Comment 1 Weston M. Price 2012-02-02 15:03:58 UTC
Currently we are not support failover/clustering in the current release of the JCA adapter due to issues upstream with the JMS client. As such, this bug should be moved out to a later release.

Comment 4 Justin Ross 2013-02-22 15:42:17 UTC
2.4/0.22 doesn't have the clustering impl involved here.

Comment 5 Weston M. Price 2013-03-12 12:07:00 UTC
This type of bug has been expressed in varying BZ's. The failover logic is implemented correctly now with the exception for HA/XA which we are working on. THis is an old bug and as such I am closing it.