Thanks Alex, Actually the service stop responding and we have seen this error in the service log.
We restart the service not Terracotta server. However this seems like a deadlock, and we don't know why is happening.
I will go a head and open a ticket in EHcache.org jira.

I was thinking may be ignoring the cache while saving the object -> cause any conflict between the 2nd level cache and the DB???

But this is what hibernate recommend if you are doing batch insert/update to avoid a memory issues???

Thanks.