I found increasing the JVM to 800meg only delayed the out of memory problems. I profiled for a while on the production machine but this only revealed a lot of stacktraces accumulating and these were eventually thrown away during a GC.
I tried disabling the memory shrinker like you suggested and so far the JVM hasn't requested an increase above it's starting allocation of 80Meg which is promising.
It's still a bit early to be 100% sure, but I was wondering what the memory shrinker does exactly? Is this the code that keeps the cache below the maxobjects setting?
Kind Regards, Matt.
--------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
