Bugzilla – Bug 4873
segfaults in RLS JNI code
Last modified: 2008-01-29 22:27:22
You need to log in before you can comment on or make changes to this bug.
I think I've seen this once before in our portal but as it seemed one off. However, we got these twice last night, when our code (i2u2 portal) was under high-than-usual load. See the attached JVM logs. Machines are based on debian 3.1
Created an attachment (id=1133) [details] JVM crash dump (1)
Created an attachment (id=1134) [details] JVM crash dump (2)
Alas I have no real idea beyond what's in the stack traces as to what causes the problem or how to recreate it reliably.
I looked at the stack traces. It is possible that it could be due to null parameters in the calls to the RLS jni client. See: http://bugzilla.globus.org/bugzilla/show_bug.cgi?id=4654 Otherwise, based on the stack trace alone, there isn't a whole lot I can tell from it. But again it could be related to 4654.
Created an attachment (id=1161) [details] hotspot error on a machine with java 1.5.0_06 this error log was produced from running the command vds-plan on a linux-rhel4-x86_64 machine with java 1.5.0_06
You don't happen to know what arguments were passed to the RLSClient (getLRCBulk operation) from the cPlanner, do you?
This bug is fixed by using the new RLS Client Java API in trunk (to be release in GT 4.2).