Forum: Building VoltDB Applications

Post: ERROR org.voltdb.utils.DBBPool - Arena 512 grew to 134479872 which is greater then t

ERROR org.voltdb.utils.DBBPool - Arena 512 grew to 134479872 which is greater then t
shetty_ritesh
May 5, 2010
Hi Tim,
I was trying to insert a huge 10 million record file. This worked fine even though at the end at approximately it reached out of memory error from jvm .
But 1 interesting thing happened. I did a fresh start and when it reached 6.5 million records i switched on another VM on the same machine. (No server runnig just switched ON) I got the below error in the logs.
May be there was a sudden spike in CPU or memory consumption of the box but looks like it affected the voltdb instance.
Regards
Ritesh
***********************************************************************************************************************
[java] [main] ERROR org.voltdb.utils.DBBPool - Arena 512 grew to 134479872 which is greater then the max of 134217728. This could signal a potential leak of ByteBuffers, an inadequately sized arena, or some other shortcoming in the network subsystem
[java] Arena 512 grew to 134479872 which is greater then the max of 134217728. This could signal a potential leak of ByteBuffers, an inadequately sized arena, or some other shortcoming in the network subsystem
[java] On Line 6,800,000
[java] Arena 256 grew to 134479872 which is greater then the max of 134217728. This could signal a potential leak of ByteBuffers, an inadequately sized arena, or some other shortcoming in the network subsystem
[java] [main] ERROR org.voltdb.utils.DBBPool - Arena 256 grew to 134479872 which is greater then the max of 134217728. This could signal a potential leak of ByteBuffers, an inadequately sized arena, or some other shortcoming in the network subsystem
***********************************************************************************************************************
re: ERROR org.voltdb.utils.DBBPool - Arena 512 grew to 134479872
tcallaghan
May 10, 2010
Ritesh,

These messages normally result from either the server or the client being "CPU starved".

VoltDB clients and servers need sufficient CPU resources to perform networking activities (pushing stored procedure invocation requests, pulling results, etc.). If this work is not getting done you will see the messages you attached.

Switching on the additional VM would certainly consume resources from the VMs currently running.

Let me know if this clears up your concern.

-Tim