WebLogic 10.3.5 cluster using Unicast for cluster communication.
The cluster has 42 servers. Periodically managed servers print the following messages in stdout:
<Feb 16, 2012 3:03:24 AM CST> <Info> <WorkManager> <BEA-002936> <maximum thread constraint ClusterMessaging is reached>
Is there any way to tune max thread constraint of ClusterMessaging?
Max Constraint needs to remain at the predefined value of 1 (if message ordering is enabled).
It is fine to be a slight queue build-up. If you are using such a large cluster, preferred solution is to use multicast communication.
If you can’t use multicast, make sure you apply the patch from “p12822180_103_Generic.zip” to ensure further unicast stability on 10.3.5, especially with large clusters of managed servers more than 20+.