Ok, so the crash is from glusterfs_handle_terminate

top = glusterfsd_ctx->active->first;

glusterfsd_ctx->active is NULL

I think having the graph traversal done inside a lock (mainly for brick
attach & detach requests, server_setvolume) might help here, Jeff?



On Mon, May 22, 2017 at 9:25 PM, Milind Changire <[email protected]>
wrote:

> backtraces are available at the job page mentioned in original mail
>
> Milind
>
> On 05/22/2017 09:22 PM, Atin Mukherjee wrote:
>
>> Do you have the backtraces?
>>
>> On Mon, 22 May 2017 at 21:20, Milind Changire <[email protected]
>> <mailto:[email protected]>> wrote:
>>
>>     Job: https://build.gluster.org/job/centos6-regression/4731/console
>>
>
_______________________________________________
Gluster-devel mailing list
[email protected]
http://lists.gluster.org/mailman/listinfo/gluster-devel

Reply via email to