user1286756 wrote:I would mention that in the bug report too, then it must be a recent bug that was added. Quite a severe one I might add.
Thanks for that although we currently use 184.108.40.206 with no problems and did use 220.127.116.11 also with no problems. I was testing 18.104.22.168 with a view to us upgrading only to get this issue.
C:\PROGRA~2\Sophos\SOPHOS~1\SOPHOS~2.DLLCertainly looks like JNI to me.
I get the same outcome ... and Linux machinesPresumably you meant a SEGSEGV. Which with JNI problem would make it deterministic and probably also some obvious pointer mistake in the JNI code. Probably not a buffer overrun and really like to be a dereference of a null pointer.
J1286756 wrote:Fixing the JNI code would seem like an obvious one to me.
So is/are there any solution(s) out there fit for production?
When I run with -Xint flag, JVM doesnt crash but of course the performance hit is profoundChanging the execution path of an application can and often does change what impact a pointer bug has in code. So not very surprising that you got a different result.
user13154783 wrote:That doesn't mean it is 'fixed' but rather that it works. For now.
I was the same problem and now it get resolved.
I downgrade my jre to 1.6.20 and increased permSize. It start working fine without error. Earlier it was crashes frequently.