Fix (APAR): PI15205 Status: Fix Release: 8.5.5.2,8.5.5.1,8.5.5 Operating System: AIX,HP-UX,IBM i,Linux,Mac OS,Solaris,Windows,z/OS Supersedes Fixes: CMVC Defect: xxxxxx Byte size of APAR: 262977 Date: 2014-06-10 Abstract: cannot connect to messaging engine after failover Description/symptom of problem: PI15205 resolves the following problem: ERROR DESCRIPTION: A messaging engine stopped on one cluster member due to database errors. The messaging engine subsequently started successfully on another cluster member. However other servers within the cell cannot connect to it. Possible symptom message: CWSIT0019E: No suitable messaging engine is available on bus that matched the specified connection properties

Traces for WLM*=all show the following exception occurring repeatedly: com.ibm.ws.cluster.selection.NoAvailableTargetExceptionImpl Within the detailed exception message, the string "removed: no CFEndPoint matching criteria" appears. LOCAL FIX: PROBLEM SUMMARY USERS AFFECTED: All users of IBM WebSphere Application Server PROBLEM DESCRIPTION: Client cannot connect to messaging engine after it fails over RECOMMENDATION: None A messaging engine stopped due to database errors. It subsequently started successfully on another cluster member. However clients cannot connect to it. Possible symptom message: CWSIT0019E: No suitable messaging engine is available on bus that matched the specified connection properties

Traces for WLM*=all show the following exception occurring repeatedly: com.ibm.ws.cluster.selection.NoAvailableTargetExceptionImpl Within the detailed exception message, the string "removed: no CFEndPoint matching criteria" appears. PROBLEM CONCLUSION: The code has been changed to correct a specific cause of this problem. This is a difficult problem to identify because there are other valid causes for the symptom message CWSIT0019E to appear. There are also other valid causes for the exception NoAvailableTargetExceptionImpl to appear in traces with the reason "removed: no CFEndPoint matching criteria". The problem occurs because an incorrect relationship forms between two objects, the RuleEtiquette and the ChannelTargetRule. It may only be possible to observe the incorrect relationship by having WLM traces active at the time the messaging engine fails over. The specific trace sequence is shown below. [3/25/14 23:06:08:106 CST] 00000024 RuleEtiquette > handleNotification Entry {CELL=PSCell1, COMPONENT=CF, NAME=InboundSecureMessaging, NODE=Node2, SERVER=MEClusterMember2, TYPE=CHANNEL} type.memento.updated [EndPointImpl$MementoImpl#-415927217{CELL=PSCell1, COMPONENT=CF, NAME=InboundSecureMessaging, NODE=Node2, SERVER=MEClusterMember2, TYPE=CHANNEL}][true::{CELL=PSCell1, COMPONENT=CF, NAME=InboundSecureMessaging, NODE=Node2, SERVER=MEClusterMember2, TYPE=CHANNEL}:false] [3/25/14 23:06:08:106 CST] 00000024 RuleEtiquette 3 working on rule com.ibm.ws.cluster.channel.ChannelTargetRule@b4338b5com.ibm.ws.c luster.channel.ChannelTargetRule@-9250769 [cfEndPointCriteria=com.ibm.wsspi.cluster.adapter.channel.Channe lSelectionCriteria@de6b082a [clusterIdentity={bus=BPM.De1.Bus, name=MECluster, type=WSAF_SIB_BUSMEMBER}] [criteria = com.ibm.ws.cluster.selection.SelectionCriteriaImpl@472ee3c0[{bus =BPM.De1.Bus, name=MECluster, type=WSAF_SIB_BUSMEMBER}:{rules.precedence=[Lcom.ibm.wsspi.clust er.selection.SelectionRule;@3e68712a, AcceptableStates=[B@b23436e3, rules.restriction=[Lcom.ibm.wsspi.cluster.selection.SelectionRul e;@8d7df22b}]] [chainName=InboundSecureMessaging] [sslRequired=false] [accessor=null] [channelFactories={null}] [3/25/14 23:06:08:106 CST] 00000024 RuleEtiquette 3 rule class name is com.ibm.ws.cluster.channel.ChannelTargetRule [3/25/14 23:06:08:106 CST] 00000024 ChannelTarget > subset Entry [{bus=BPM.De1.Bus, name=MECluster.000-BPM.De1.Bus, subnet=, type=WSAF_SIB_ME_LONG_NAME}] [3/25/14 23:06:08:106 CST] 00000024 ChannelTarget < subset Exit [{bus=BPM.De1.Bus, name=MECluster.000-BPM.De1.Bus, subnet=, type=WSAF_SIB_ME_LONG_NAME}] The key points in this trace are: 1) Entry to RuleEtiquette.handleNotification() with one of the arguments being type.memento.updated. followed by 2) Entry to ChannelTargetRule.subset() with a cluster identity that does not represent an application server (i.e. something like [{CELLNAME=PSCell1, MEMBERNAME=MEClusterMember2, NODENAME=Node2}]). This shows the incorrect relationship exists between the RuleEtiquette and ChannelTargetRule objects. The fix for this APAR is currently targeted for inclusion in fix pack 8.5.5.3. Please refer to the Recommended Updates page for delivery information: http://www.ibm.com/support/docview.wss?rs=180&uid=swg27004980 Directions to apply fix: Fix applies to Editions: Release 8.5 __ Application Server (Express or BASE) X_ Network Deployment (ND) __ Edge Components __ Developer Install Fix to all WebSphere installations unless special instructions are included below. Special Instructions: None NOTE: The user must: * Logged in with the same authority level when unpacking a fix, fix pack or refresh pack. * Be at V1.4.3 or newer of the Installation Manager. Certain iFixes may require a newer version of the Installation Manager and the Installation Manager will inform you during the installation process if a newer version is required. The IBM Information Center can provide details, if needed, on the use of the Installation Manager to apply the iFixes. http://publib.boulder.ibm.com/infocenter/install/v1r4/index.jsp. Shutdown WebSphere Application Server before applying the iFixes. Restart WebSphere Application Server after applying the iFixes. Directions to remove fix: The IBM Information Center can provide details, if needed, on the use of the Installation Manager to remove the iFixes. http://publib.boulder.ibm.com/infocenter/install/v1r4/index.jsp. Shutdown WebSphere Application Server before removing the iFixes. Restart WebSphere Application Server after removing the iFixes. Directions to re-apply fix: 1) Shutdown WebSphere Application Server. 2) Follow the Fix instructions to apply the fix. 3) Restart WebSphere Application Server. Additional Information: