I've done some further testing, and it seems that the SIGSEGV crashes are coming from Karaf code - the core dump from one of them references hawtjni.c, and Karaf is the only component of our project which uses this. Tested updating to 2.2.10, and the problem persists. Didn't notice the problem on 2.3.0, but that (along with OSGi 4.3) introduces a lot of new issues with our project (and we really want to avoid a major Karaf version change for our next release).
(The original issues I had with upgrading turned out to be due to old code in my Maven repository). With that said, does anybody have any idea as to how to work around the issue in Karaf 2.2.x? The issue doesn't seem to happen as much (or at all) on 32-bit/non-Oracle JVMs, though that's not really a viable long-term solution. Tim FYI, here is the error message again, in case anybody missed... # A fatal error has been detected by the Java Runtime Environment: # # SIGSEGV (0xb) at pc=0x00007f5ebd3fc54e, pid=23273, tid=140044176189184 # # JRE version: 7.0_09-b05 # Java VM: Java HotSpot(TM) 64-Bit Server VM (23.5-b02 mixed mode linux-amd64 compressed oops) # Problematic frame: # C [libc.so.6+0x8054e][thread 140044020991744 also had an error] -- View this message in context: http://karaf.922171.n3.nabble.com/SIGSEGV-crash-issues-with-Karaf-OSGi-on-Linux-tp4027102p4027218.html Sent from the Karaf - User mailing list archive at Nabble.com.
