Improving the Performance of Connect Meeting Clustered Fail-over
To improve the performance of Connect Meeting fail-over in a cluster, make the following changes to the custom.ini file and cycle the FMS and Connect services or reboot the connect servers:
The installation documentation for clustering originally prescribed setting the MEETING_TIMEOUT variable to 0 as shown:
#For Connect 8&9 Cluster fail-over add the following
We have discovered that in versions of Connect 9 prior to 9.2, setting the MEETING_TIMEOUT variable to 1 insures that a meeting will only have one active session while setting the MEETING_TIMEOUT variable to 0 opens the possibility (although rare) of a meeting having more than one active session. The result is a bit odd. Everyone could be happily in an ongoing Connect meeting while one attendee is alone in what appears to be the same meeting. This is usually triggered by an some form interruption in the isolated users meeting session such as from a brief outage in the network connection of the isolated user.
The recommended custom.ini setting for Connect 9 servers prior to version 9.2 looks as follows:
#For Connect 9 Cluster fail-over add the following
Save the custom.ini after making the change and cycle the Flash Management Sever and Adobe Connect Server services.
In Connect version 9.2, the meeting timeout session may be set back to 0 as we have made the appropriate code changes to insure the functionality works as originally intended even if a user experiences network interruptions and must reconnect to the meeting.