What kind of eviction are you using? Overflow-to-disk or destroy? If overflow, keep in mind that keys and entries are always in memory. Only values are overflowed to disk. I guess we'd need to see your GemFire stats and log files to see whats going on. Can you make those available?
Distributed System shutDown Hook
To evoid loading id already cached I put 20 records in a map and then cached the map....
System goes faster and suddently after 800.000 records casched (x20) I get the following error, but without block the pc or slow...
Exception in thread "UDP Incoming Message Handler" Exception in thread "main" java.lang.OutOfMemoryError: Java heap space
java.lang.OutOfMemoryError: Java heap space
Exception in thread "TimeScheduler.Thread" java.lang.OutOfMemoryError: Java heap space
Exception: java.lang.OutOfMemoryError thrown from the UncaughtExceptionHandler in thread "Thread-2 StatSampler"
Exception: java.lang.OutOfMemoryError thrown from the UncaughtExceptionHandler in thread "UDP mcast receiver"
Exception: java.lang.OutOfMemoryError thrown from the UncaughtExceptionHandler in thread "Distributed system shutdown hook"
I followed the instruction from this post about the xml setting:
and this post is the result
keep in mind that they are more then 100Millions of trandactions to cache....
Please sign in to leave a comment.
this is the xml
<lru-entry-count maximum="700000" action="overflow-to-disk"/>
but I tried both overflow and destroy.....
I would like a thing that after n records everything will be casched, otherwise it wouldnt have any sense to build a structure like gemfire.
So the only way is to give arguments to VM like -Xmx ?
As regarding the log later I ll upload it... Thanks