[Gluster-devel] glusterd 3.8dev crash - core dump

Anand Nekkunti anekkunt at redhat.com
Tue May 12 05:09:54 UTC 2015


     This bt similar to 
https://bugzilla.redhat.com/show_bug.cgi?id=1209461 . This crash due to 
race between clean up thread and rpc event thread .It is known issue , 
It doesn't have any serious impact to functionality apart from creating 
core  dump file and log message.

for more info :
https://bugzilla.redhat.com/show_bug.cgi?id=1209461

I have sent patch for rpc listener clean up 
(http://review.gluster.org/10197 ) , but glusterd fini()  function is 
not invoking  before exit(0)  due to code comment in cleanup_and_exit() 
( no idea why it is commented ).

void cleanup_and_exit (int signum)
{
          ------
          -------
           glusterfs_pidfile_cleanup (ctx);

         exit (0);
#if 0
         /* TODO: Properly do cleanup_and_exit(), with synchronization */
         if (ctx->mgmt) {
                 /* cleanup the saved-frames before last unref */
                 rpc_clnt_connection_cleanup (&ctx->mgmt->conn);
                 rpc_clnt_unref (ctx->mgmt);
         }

         /* call fini() of each xlator */
         trav = NULL;
         if (ctx->active)
                 trav = ctx->active->top;
         while (trav) {
                 if (trav->fini) {
                         THIS = trav;
                         trav->fini (trav);
                 }
                 trav = trav->next;
         }
#endif
}


Regards
Anand.N

On 05/12/2015 09:09 AM, Christopher Pereira wrote:
> On 10-05-2015 6:26, Niels de  Vos wrote:
 >> On Sat, May 09, 2015 at 06:34:55AM -0300, Christopher Pereira
 >> wrote:
 >>> Core was generated by `glusterd --xlator-option *.upgrade=on
 >>> -N'. Program terminated with signal 11, Segmentation fault. #0
 >>> 0x00007f489c747c3b in ?? ()
 >>>
 >>> [...]
 >
 > Bug reported here:
 > https://bugzilla.redhat.com/show_bug.cgi?id=1220623
 >
 > _______________________________________________ Gluster-devel mailing
 > list Gluster-devel at gluster.org
 > http://www.gluster.org/mailman/listinfo/gluster-devel


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.gluster.org/pipermail/gluster-devel/attachments/20150512/17ff5896/attachment-0001.html>


More information about the Gluster-devel mailing list