On Thu, Jul 13, 2017, at 01:10 PM, Pranith Kumar Karampuri wrote:
I just observed that https://build.gluster.org/job/centos6-regression/5433/consoleFull failed because of this .t failure.
Maybe we need to upgrade the version of gdb on these machines, because it didn't seem able to get the backtrace but the version I'm using (7.12.50) did just fine. The first few frames of thread 1 above the signal handler look like this.
#8 0x00007f9b642d8872 in fini_db (_conn_node=0x7f9b528d2e50)
at /home/jenkins/root/workspace/centos6-regression/libglusterfs/src/gfdb/gfdb_dat
a_store.c:326
#9 0x00007f9b644ffb8e in notify (this=0x7f9b524795e0, event=9, data="">
at /home/jenkins/root/workspace/centos6-regression/xlators/features/changetimerec
order/src/changetimerecorder.c:2313
#10 0x00007f9b7280c04f in xlator_notify (xl=0x7f9b524795e0, event=9, data="">
b20)
at /home/jenkins/root/workspace/centos6-regression/libglusterfs/src/xlator.c:566
#11 0x00007f9b728cf1ec in default_notify (this=0x7f9b5247bb20, event=9, data="">
47cf90) at defaults.c:3151
#12 0x00007f9b5fded58d in notify (this=0x7f9b5247bb20, event=9, data="">
at /home/jenkins/root/workspace/centos6-regression/xlators/features/changelog/src
/changelog.c:2307
Based on that, I think I'd start looking at changetimerecorder/libgfdb first. IIRC we've had some init/fini races there before.
_______________________________________________ Gluster-devel mailing list Gluster-devel@xxxxxxxxxxx http://lists.gluster.org/mailman/listinfo/gluster-devel