[Gluster-devel] tests/bugs/snapshot/bug-1109889.t - snapd crash

Raghavendra G raghavendra at gluster.com
Mon Jul 6 04:55:01 UTC 2015


<server_setvolume>

         */
        if (op_ret && !xl) {
                /* We would have set the xl_private of the transport to
the

                 * @conn. But if we have put the connection i.e shutting
down

                 * the connection, then we should set xl_private to NULL as
it
                 * would be pointing to a freed memory and would segfault
when
                 * accessed upon getting
DISCONNECT.

                 */
                gf_client_put (client, NULL);
                req->trans->xl_private = NULL;
        }

</server_setvolume>

The crash is in gf_client_put.  Code in gf_client_put reveals that client
is dereferenced without NULL check. I am suspecting that this crash
might've been uncovered/caused by [1] which fails any setvolume requests
before server graph initialization (in which case client is NULL). Will
send out a patch.

[1] http://review.gluster.org/11490

On Fri, Jul 3, 2015 at 6:02 PM, Raghavendra Bhat <rabhat at redhat.com> wrote:

> On 07/03/2015 03:37 PM, Atin Mukherjee wrote:
>
>>
>> http://build.gluster.org/job/rackspace-regression-2GB-triggered/11898/consoleFull
>> has caused a crash in snapd with the following bt:
>>
>
> This seem to have crashed in server_setvolume (i.e. before the graph could
> be properly made available for i/o. snapview-server xlator is yet to come
> into the picture). But still I will try to reproduce it on my local setup
> and see what might be causing this.
>
>
> Regards,
> Raghavendra Bhat
>
>
>
>> #0  0x00007f11e2ed3ded in gf_client_put (client=0x0, detached=0x0)
>>      at
>>
>> /home/jenkins/root/workspace/rackspace-regression-2GB-triggered/libglusterfs/src/client_t.c:294
>> #1  0x00007f11d4eeac96 in server_setvolume (req=0x7f11c000195c)
>>      at
>>
>> /home/jenkins/root/workspace/rackspace-regression-2GB-triggered/xlators/protocol/server/src/server-handshake.c:710
>> #2  0x00007f11e2c1e05c in rpcsvc_handle_rpc_call (svc=0x7f11d001b160,
>> trans=0x7f11c0000ac0, msg=0x7f11c0001810)
>>      at
>>
>> /home/jenkins/root/workspace/rackspace-regression-2GB-triggered/rpc/rpc-lib/src/rpcsvc.c:698
>> #3  0x00007f11e2c1e3cf in rpcsvc_notify (trans=0x7f11c0000ac0,
>> mydata=0x7f11d001b160, event=RPC_TRANSPORT_MSG_RECEIVED,
>>      data=0x7f11c0001810) at
>>
>> /home/jenkins/root/workspace/rackspace-regression-2GB-triggered/rpc/rpc-lib/src/rpcsvc.c:792
>> #4  0x00007f11e2c23ad7 in rpc_transport_notify (this=0x7f11c0000ac0,
>> event=RPC_TRANSPORT_MSG_RECEIVED, data=0x7f11c0001810)
>>      at
>>
>> /home/jenkins/root/workspace/rackspace-regression-2GB-triggered/rpc/rpc-lib/src/rpc-transport.c:538
>> #5  0x00007f11d841787b in socket_event_poll_in (this=0x7f11c0000ac0)
>>      at
>>
>> /home/jenkins/root/workspace/rackspace-regression-2GB-triggered/rpc/rpc-transport/socket/src/socket.c:2285
>> #6  0x00007f11d8417dd1 in socket_event_handler (fd=13, idx=3,
>> data=0x7f11c0000ac0, poll_in=1, poll_out=0, poll_err=0)
>>      at
>>
>> /home/jenkins/root/workspace/rackspace-regression-2GB-triggered/rpc/rpc-transport/socket/src/socket.c:2398
>> #7  0x00007f11e2ed79ec in event_dispatch_epoll_handler
>> (event_pool=0x13bb040, event=0x7f11d4eb9e70)
>>      at
>>
>> /home/jenkins/root/workspace/rackspace-regression-2GB-triggered/libglusterfs/src/event-epoll.c:570
>> #8  0x00007f11e2ed7dda in event_dispatch_epoll_worker
>> (data=0x7f11d000dc10)
>>      at
>>
>> /home/jenkins/root/workspace/rackspace-regression-2GB-triggered/libglusterfs/src/event-epoll.c:673
>> #9  0x00007f11e213e9d1 in start_thread () from ./lib64/libpthread.so.0
>> #10 0x00007f11e1aa88fd in clone () from ./lib64/libc.so.6
>>
>>
> _______________________________________________
> Gluster-devel mailing list
> Gluster-devel at gluster.org
> http://www.gluster.org/mailman/listinfo/gluster-devel
>



-- 
Raghavendra G
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.gluster.org/pipermail/gluster-devel/attachments/20150706/73dbec98/attachment.html>


More information about the Gluster-devel mailing list