[GE users] Crashing dbwriter

beatrubi beat at 0x1b.ch
Fri Mar 12 08:04:28 GMT 2010


Hello!

I manage a cluster of ~260 nodes with full reporting using GE and ARCO
6.2u3, JDK 1.6.0_14, MySQL 5.0.26 and mysql-connector-java-5.1.10.
Approximately every two weeks the dbriter is crashing:

12/03/2010 02:36:31|master|ngDBWriter$StatisticThread.run|E|Unknown error:
java.lang.OutOfMemoryError: Java heap space

I increased the heapspace to 512 Mbytes, setting JVMARGS at the beginning of
the script was not working:

---8<---
   if [ "$JVMARGS" = "" ]; then
      JVMARGS="-Xms512m -Xmx512m -classpath $CP"
   else
      JVMARGS="$JVMARGS -Xms512m -Xmx512m -classpath $CP"
   fi 
---8<---

Not sure if this helps or if it's simple delaying the crash. At least it
should help for two months, the default heapsize seems to be 64m :-)

Did I hit some known memory leak?

Beat

-- 
     \|/                           Beat Rubischon <beat at 0x1b.ch>
   ( 0^0 )                             http://www.0x1b.ch/~beat/
oOO--(_)--OOo---------------------------------------------------
Meine Erlebnisse, Gedanken und Traeume: http://www.0x1b.ch/blog/

------------------------------------------------------
http://gridengine.sunsource.net/ds/viewMessage.do?dsForumId=38&dsMessageId=248123

To unsubscribe from this discussion, e-mail: [users-unsubscribe at gridengine.sunsource.net].



More information about the gridengine-users mailing list