KP,
One way to view relevant information from the core would be as follow,
1) before loading the core you need to tell gdb to look for shared objects in a
different path,
- using 'set solib-search-path'
- For us this translates to 'set solib-search-path
./lib:./lib/glusterfs:./lib/glusterfs/3.5git:./lib/glusterfs/3.5git/auth:./lib/glusterfs/3.5git/rpc-transport:./lib/glusterfs/3.5git/xlator:./lib/glusterfs/3.5git/xlator/cluster:./lib/glusterfs/3.5git/xlator/debug:./lib/glusterfs/3.5git/xlator/encryption:./lib/glusterfs/3.5git/xlator/features:./lib/glusterfs/3.5git/xlator/features/glupy:./lib/glusterfs/3.5git/xlator/mgmt:./lib/glusterfs/3.5git/xlator/mount:./lib/glusterfs/3.5git/xlator/nfs:./lib/glusterfs/3.5git/xlator/performance:./lib/glusterfs/3.5git/xlator/protocol:./lib/glusterfs/3.5git/xlator/storage:./lib/glusterfs/3.5git/xlator/system:./lib/glusterfs/3.5git/xlator/testing:./lib/glusterfs/3.5git/xlator/testing/features:./lib/glusterfs/3.5git/xlator/testing/performance:./lib/ocf:./lib/ocf/resource.d:./lib/ocf/resource.d/glusterfs:./lib/pkgconfig:./lib/python2.6:./lib/python2.6/site-packages:./lib/python2.6/site-packages/gluster'
I know the above is more than what you need to debug glusterd (which is the
current cored process), but as you do not know which binary crashed, better to
add all from a more generic standpoint.
This would be done starting a vanilla gdb session and entering the command
there before opening the core.
NOTE: here my pwd (or resolution of . in ./lib/--- is where I untarred the core
tarball and /untar location/build/install/ so that it searches all shared
objects from the jenkins system that was bundled into the tar file.
2) load the core file in gdb now as follows (or other methods known),
- (gdb) target exec ./sbin/glusterd
- (gdb) target core ./cores/core.1077
- (gdb) bt, other commands and enjoy for the most part!
Shyam
- Original Message -
From: Krishnan Parthasarathi kpart...@redhat.com
To: Pranith Kumar Karampuri pkara...@redhat.com
Cc: Gluster Devel gluster-devel@gluster.org
Sent: Thursday, June 19, 2014 5:52:42 AM
Subject: [Gluster-devel] Regarding regression failure in
rackspace-regression-2GB machine
Pranith,
The core's backtrace [1] is not 'analysable'. It doesn't show function names
and displays ()? for all the frames across all threads. It would be helpful
if we had the glusterd logs corresponding to cluster.rc setup. These logs
are missing too.
thanks,
Krish
[1] - glusterd core file can be found here -
http://build.gluster.org/job/rackspace-regression-2GB/250/consoleFull
___
Gluster-devel mailing list
Gluster-devel@gluster.org
http://supercolony.gluster.org/mailman/listinfo/gluster-devel
___
Gluster-devel mailing list
Gluster-devel@gluster.org
http://supercolony.gluster.org/mailman/listinfo/gluster-devel