We're experiencing the problem that seems to be documented in the tagged question, but the solution provided has consequences in VMWare where disabling memory ballooning causes performance issues for the physical host, thus affecting other unrelated servers.
Is there any other suggestion that can be offered? Would increased RAM help (even though it's the CPU that's pegged, not the memory)? Adding another CPU (or will it run away with both)?
As best I understand this has something to do with Java garbage collection - is there any configuration change I can make or process I can trigger to overcome the more pinpointed issue?
