Table Of Contents
Cisco Unity 4.0(5) Engineering Special 32
Published December14, 2005
MWI resync process is lengthy after Exchange cluster failover.
Problem Description and Resolution
When the Exchange Cluster is failed over from NodeA to NodeB, Unity releases MAPI interfaces that had previously been established on NodeA. During this release, MAPI attempts to communicate with the EMSMDB store provider that had previously existed on NodeA. It takes a while at the network layer for MAPI to "figure out" it should contact NodeB since NodeA is no longer the Node in the cluster for Exchange. Below is an example of diagnostic information from Unity's ExchangeMonitor component (which resides in the AvMsgStoreMonitorSvr service) showing this delay. It takes ~27 seconds for the request of the release.
18:38:03:852 (AvDiagnostics_MC,645,ExchangeMonitor,11) [Thread 6160]CAvNotifyContext Releasing mailbox cn=TestUser cn=Recipients ou=First Administrative Group o=First Organization
18:38:30:164 (AvDiagnostics_MC,645,ExchangeMonitor,11) [Thread 6160]CAvNotifyContext Released mailbox cn=TestUser cn=Recipients ou=First Administrative Group o=First Organization
With this ES, the ExchangeMonitor will release the MAPI interfaces on a thread that is not used for the resync process.
Special Installation Instructions
New and Changed Behavior
© 2005 Cisco Systems, Inc. All rights reserved.