we use Infinispan 5.3.0 final in our application which runs in a 2 machine cluster. We have multiple caches about 10 with different data in it and with the Infinispan Configuration as attached to this post. Attached to this post is also the jgroups config we use for our application.
Because we have the problem that all 2 days there are SuspectExceptions occuring on the servers, my first question will be how to handle such Exceptions when using Infinispan. And also maybe someone can tell us if there is a mistake in the jgroups which causes the problem. Maybe something has to be changes under heavy load ? What is also a big problem is that after the SuspectExceptions occured there are some lock not released by Infinispan (sometimes only on 1 machine). This leads to TimeoutExceptions every time the lock has to be acquired. Our only workaround since now is to start a new EJB timer and unlock these specific lockIds in the cache. But the question here is what is the best pratice to handle this, because nothing could be found regarding such a situation until yet. Maybe as already metioned there is something completly wrong with our configuration, but I hope you can help us with that?