Resolution:
Description:
===========
Configuring fault tolerant Coherence cache  (Cache Servers started on separate machines).

Environment:
===========
BusinessEvents 3.0.2
All Operating Systems

Symptoms:
=========
Lose cache data handled by CacheServers , when one of the machines goes down.

Cause:
======
Copy of the Cache backup (backupcount > 1)  not available on one of the other machines (configuration issue).

Resolution:
========
Coherence automatically handles the distribution and recovery of entities in the cluster. Coherence also does its best to ensure "a backup partition will not reside on the same physical machine as the primary partition".

The parameters below should ensure that copies are backed up on different nodes. At a minimum, tangosol.coherence.machine needs to be set in each node.

java.property.tangosol.coherence.cluster
java.property.tangosol.coherence.site
java.property.tangosol.coherence.rack
java.property.tangosol.coherence.machine

Users can read more about these parameters at;
http://wiki.tangosol.com/display/COH32UG/member-identity

Example:

# backup count has to be set at least to 1
java.property.tangosol.coherence.machine=machine1
java.property.tangosol.coherence.site= machine1
java.property.tangosol.coherence.rack= machine1
java.property.be.engine.limited.cache.backup.count=1
java.property.objectTable.backup.count=1
java.property.tangosol.coherence.distributed.backupcount=1
# set minCacheServers to the number of CacheServers started on a single machine
be.engine.cluster.minCacheServers=3


NOTE:
1.)  We recommend to run some tests to confirm Coherence is working as expected.

2.) You have to start the same number of CacheServers on each node.

Configuring fault tolerant Coherence cache(Cache Servers started on separate machines).