PM21983: APPLICATION SERVER ATTEMPTS TO SEND SIP REQUESTS TO UNREACHABLE PROXY.

APAR status

• Closed as program error.

Error description

• Application Server detects proxy down via SNOD - yet still
  sending out via down proxy.
  
  WebSphere Application Server Network Deployment edition
  V7.0.0.11 Feature Pack for Communications Enabled Applications
  (CEA) V1.0.0.5.
  
  We are doing testing where we bring down the switch for the
  Session Initiation Protocol (SIP) signaling network on the
  first bladecenter.  Looking at an application server on the
  second bladecenter, we noticed that the log DOES show that
  SNOD is working, that it knows that the proxy is down, but
  it still round robins between the two proxies!
  [8/31/10 22:06:21:252 UTC] 00000046 HeartbeatMoni 3
  HeartbeatMonitor
  timeout [xx.xx.xx.xx:-1/TCP] [3]
  [8/31/10 22:06:21:252 UTC] 00000046 HeartbeatMoni W
  HeartbeatMonitor
  CWSCT0362I: Network heartbeat limit exceeded with SIP proxy
  xx.xx.xx.xx:-1/TCP. 3 heartbeats missed.
  [8/31/10 22:06:21:253 UTC] 00000046 HeartbeatMoni 3
  HeartbeatMonitor isNetworkDown lost touch with [1] out of [2]
  proxies
  
  Note:- xx.xx.xx.xx is an actual IP address
  

Local fix

Problem summary

• ****************************************************************
  * USERS AFFECTED:  All IBM WebSphere Application Server Feature
  *                  Pack for Communications Enabled Applications
  *                  (CEA) users                                 *
  ****************************************************************
  * PROBLEM DESCRIPTION: On a cluster topology with multiple     *
  *                      SIP Proxies, in which the SIP network   *
  *                      is separated from the management        *
  *                      network, if only part of the proxies    *
  *                      lost SIP connection with the servers,   *
  *                      this can lead to call failures.         *
  ****************************************************************
  * RECOMMENDATION:                                              *
  ****************************************************************
  When the SIP Network Outage Detection (SNOD) feature is
  enabled, the Session Initiation Protocol (SIP) proxies
  periodically send SIP KEEPALIVE messages to the servers. If
  all proxies stopped sending KEEPALIVE to a server, it will
  restart itself to initiate sessions failover to its backup
  peer. If there is at least one proxy still connected with SIP
  to the server, the server will not restart itself. The problem
  was that the SIP server still counted the disconnected proxies
  in its proxy round-robin selection for sending outgoing
  messages to, which led to error in sending some of the SIP
  messages in the dialog.
  

Problem conclusion

• The SIP server will remove a proxy that fails to send
  KEEPALIVE messages from its round-robin proxies selection
  table, and will add it back if the connection was recovered.
  
  The fix for this APAR is currently targeted for inclusion in
  fix pack for FEP CEA 1.0.0.7. Please refer to the Recommended
  Updates page for delivery information:
  http://www.ibm.com/support/docview.wss?rs=180&uid=swg27004980
  

Temporary fix

Comments

APAR Information

• APAR is sysrouted FROM one or more of the following:

• APAR is sysrouted TO one or more of the following:

Fix information

• Fixed component name

  CEA FEATUREPACK

• Fixed component ID

  5724J0855

Applicable component levels

• R700 PSY

     UP