No updates arrives in a GMS when POA runs in a cluster

  • 7000663
  • 12-Jun-2008
  • 10-Dec-2013

Environment

Novell GroupWise 7, SP3
GroupWise Mobile Sever 2.0x

Situation

If GroupWise post office runs in a cluster, no updates are synchronized to GMS mailboxes.

Resolution

In a synchronization technology that was implemented in a GMS, it is a GroupWise POA that sends a kind of a slap to a GMS server whenever any of monitored items get changed within any existing / registered GroupWise mailbox. Then in turn, a GMS server makes a full request for a data from a mailbox which it was notified about. But when POA sends a slap to a GMS, in case of cluster environment, it is sent from a host IP address and not from assigned secondary IP address (a cluster resource IP address).
This issue has been reported to a development. A current workaround is not to use "Bind exclusively ..." option on POA | Network Address tab. In this case, POA listens on both, a host and assigned cluster resource IP addresses. A GMS server sends over a time a sanitary request (by default each 30 minutes) that pools GroupWise post offices for updates. This GroupWise Poller process also uses incorrectly host IP address but since PO now does listen on both IP addresses, it replies to a GMS server which updates GMS mailboxes / mobile devices.

Additional Information

This problem can be identified when taking a trace of a communication between a GroupWise POA (in a cluster) and a GMS server. GroupWise slap uses by default a port #8191. GroupWise Poller uses a SOAP port #, a default is 7191 but it can be changed from ConsoleOne.