[Gluster-devel] New glusterd crash, at least consistent on my laptop

[Atin Mukherjee]

On 05/03/2015 11:32 PM, Vijay Bellur wrote:
> On 05/03/2015 10:25 PM, Atin Mukherjee wrote:
>>
>>
>> On 05/03/2015 05:25 PM, Pranith Kumar Karampuri wrote:
>>> Execute the following command on replicate volume:
>>> root at pranithk-laptop - ~
>>> 17:23:02 :( ⚡ gluster v set r2 cluster.client-log-level 0
>>> Connection failed. Please check if gluster daemon is operational.
>>>
>>> Program terminated with signal SIGSEGV, Segmentation fault.
>>> #0 0x00000038e480c860 in pthread_spin_lock () from
>>> /lib64/libpthread.so.0
>>> (gdb) bt
>>> #0 0x00000038e480c860 in pthread_spin_lock () from
>>> /lib64/libpthread.so.0
>>> #1 0x00007f80cb27ca32 in __gf_free (free_ptr=0x7f80a81ad760) at
>>> mem-pool.c:303
>>> #2 0x00007f80c197a13a in gd_sync_task_begin (op_ctx=0x7f80a800358c,
>>> req=0x7f80b0005ffc) at glusterd-syncop.c:1826
>>> #3 0x00007f80c197a1d0 in glusterd_op_begin_synctask (req=0x7f80b0005ffc,
>>> op=GD_OP_SET_VOLUME, dict=0x7f80a800358c) at glusterd-syncop.c:1846
>>> #4 0x00007f80c18dca39 in __glusterd_handle_set_volume
>>> (req=0x7f80b0005ffc) at glusterd-handler.c:1871
>>> #5 0x00007f80c18d784a in glusterd_big_locked_handler
>>> (req=0x7f80b0005ffc, actor_fn=0x7f80c18dc532
>>> <__glusterd_handle_set_volume>)
>>> at glusterd-handler.c:83
>>> #6 0x00007f80c18dcb2c in glusterd_handle_set_volume (req=0x7f80b0005ffc)
>>> at glusterd-handler.c:1893
>>> #7 0x00007f80cb28ef11 in synctask_wrap (old_task=0x7f80b0006b00) at
>>> syncop.c:375
>>> #8 0x00000038e4047a00 in ?? () from /lib64/libc.so.6
>>> #9 0x0000000000000000 in ?? ()
>> I believe you were running a debug build. We have already hit this crash
>> few days ago and http://review.gluster.org/#/c/10238/ should solve it.
>>
>> Vijay,
>>
>> Can you please take a look at the patch and get it merged. Couple of +1s
>> are already there.   
> 
> I am looking at Jeff's more generic patch [1] to help solve this class
> of problems. More reviews on that would be helpful.
Ohh, yes, I forgot about this patch which Jeff has posted few days ago.
Jeff,

Can we abandon http://review.gluster.org/#/c/10238/ considering [1] is
the generic fix.

~Atin
> 
> -Vijay
> 
> [1] http://review.gluster.org/#/c/10417/
> 

-- 
~Atin