2 Replies Latest reply on Mar 17, 2003 2:42 PM by Sacha Labourey

    snmp crashing jboss & jvm

    nick mills Newbie

      This is a tough one. Not really sure if clustering is the correct forum - because the origin of the problem is unclear.
      We have a lot of legacy monitoring tools that use snmp, including MRTG and some homegrown tools.
      Our environment is 4 clustered jboss-3.0.6 nodes running on j2sdk1.4.1 on 4 Redhat 7.3 servers. We have a lot of network traffic, throughput is up to 3Mb/sec.

      When we run the monitoring tools we sometimes get nasty JVM crashes along with a hs_err_pid2091.log type files, which I have attached. There is no error output to the jboss logs. Using the "ps" command there are no leftover jboss processes, I guess this is because the JVM itself has crashed.

      I don't know much about snmp, but am aware it uses UDP (as does the clustered jboss).

      Could there be data packets from our snmp tools somehow infiltrating into jboss and crashing our JVM's?

      Any suggestions gratefully accepted.