Solved [Solved] openjdk-7 segfaults with 10-STABLE?

E

ecd

I've upgraded from 9-STABLE recently and have been repeatedly getting SIGSEGVs while using java/openjdk7. The segfaults happen randomly but regularly; I receive them 3/5 times when running Maven mvn clean install on a project and my IntelliJ session crashes sometimes once every ten minutes or once per day. When using 9-STABLE with the same version of openjdk7, I did not have these issues. I've also updated my kernel/world to a more recent 10-STABLE to no avail.

I have plenty of core dumps and hs_err_pid files. It does not seem that the thread dumps are the same each time:
Code: 
Stack: [0x00007fffff1f5000,0x00007fffff2f5000],  sp=0x00007fffff2f3a20,  free space=1018k
Native frames: (J=compiled Java code, j=interpreted, Vv=VM code, C=native code)
C  [libc.so.7+0xa4f7d]  sbrk+0x7f2d
C  [libc.so.7+0xa446d]  sbrk+0x741d
C  [libc.so.7+0xa116f]  sbrk+0x411f
C  [libc.so.7+0xa145b]  sbrk+0x440b
C  [libc.so.7+0xac59c]  __free+0x37c
V  [libjvm.so+0x211db8]  +0x90a00
V  [libjvm.so+0x211e44]  +0x90a8c
V  [libjvm.so+0x38ba0f]  +0x20a657
V  [libjvm.so+0x38c37d]  +0x20afc5
V  [libjvm.so+0x7d9c30]  JVM_handle_bsd_signal+0x116140
V  [libjvm.so+0x6be162]  JVM_FindSignal+0x176c82
C  [libthr.so.3+0x94a4]  pthread_create+0x764

Code: 
Stack: [0x00007fffff0f4000,0x00007fffff1f4000],  sp=0x00007fffff1f0c38,  free space=1011k
Native frames: (J=compiled Java code, j=interpreted, Vv=VM code, C=native code)
C  0x000000083a2e72f8
V  [libjvm.so+0x64793d]  JVM_FindSignal+0x10045d
V  [libjvm.so+0x37da5a]  +0x1fc6a2
V  [libjvm.so+0x3861f4]  +0x204e3c
V  [libjvm.so+0x302980]  +0x1815c8
V  [libjvm.so+0x38b7ac]  +0x20a3f4
V  [libjvm.so+0x38c37d]  +0x20afc5
V  [libjvm.so+0x7d9c30]  JVM_handle_bsd_signal+0x116140
V  [libjvm.so+0x6be162]  JVM_FindSignal+0x176c82
C  [libthr.so.3+0x94a4]  pthread_create+0x764

My system information:
Code: 
$ uname -a
FreeBSD qk.local 10.0-STABLE FreeBSD 10.0-STABLE #2 r261350: Sat Feb  1 11:36:21 CST 2014     root@qk.local:/usr/obj/usr/src/sys/GENERIC  amd64
$ pkg version -e openjdk
openjdk-7.25.15_2,1

Is anybody else seeing this? Or have any suggestions?
 
Re: openjdk-7 segfaults with 10-STABLE?

This is a known issue. It has been reported to the developers via the FreeBSD mailing lists and is being worked on. If you want to follow the progress, you can browse the freebsd-java@freebsd.org mailing list.
 
Re: openjdk-7 segfaults with 10-STABLE?

trh411 said:
This is a known issue. It has been reported to the developers via the FreeBSD mailing lists and is being worked. If you want to follow the progress, you can browse the freebsd-java@freebsd.org mailing list.
Click to expand...
I've been trying to follow this but I have yet to find definitive information.

-"7.25.15_2 or earlier should be safe." [1]
Click to expand...
I'm using this version

"If your machine can /build/ the latest OpenJDK without panicing (which involves extensive use of Java to compile itself) then you're OK to deploy that version to run your web applications" [1]
Click to expand...
I had no difficulty compiling the openjdk7 port.

"FreeBSD 11-CURRENT (r259951), 10-STABLE (r260081), 10.0-RELEASE-rc4 (r260122) and 9-STABLE (r260082) have been patched" [1]
Click to expand...
My 10-STABLE version (r261350) is one month newer than r260081.

[1] http://lists.freebsd.org/pipermail/free ... 10478.html
 
Re: openjdk-7 segfaults with 10-STABLE?

OP here.

I have yet to resolve this. I've updated my kernel/world to 10-STABLE r261660 and rebuilt java/openjdk7 with no change. The options I'm looking at trying next:
  • use Java only in VirtualBox;
  • go back to 10-RELEASE;
  • go back to 9-STABLE.
 
Re: openjdk-7 segfaults with 10-STABLE?

A couple more attempts:
  • java/openjdk-7.51.13,1 was just released to ports. I installed this, but it still crashes as before.
  • java/linux-sun-jdk17-7.51 also had regular SIGSEGVs though it also failed in new ways as well with mvn clean install giving
    Code: 
    VM crash or System.exit called ?
 
[SOLVED] Re: openjdk-7 segfaults with 10-STABLE?

OP here

I had given up on this and been using VirtualBox with a Linux guest, but I decided to give it another try. Today I upgraded to r262928 and it's now working without issues. I don't know which specific revision fixed it.
 
You must log in or register to reply here.
Back
Top