[Gluster-users] GlusterFS 3.3 span volume, op(RENAME(8)) fails, "Transport endpoint is not connected"

Amar Tumballi amarts at redhat.com
Thu Jun 14 11:38:31 UTC 2012

> 1. What is a "process state dump" and how would I get it?

'state dump' is the dump of state the process at the time. It can give 
details like how many calls pending, how much memory used, where, etc etc.

You can get it by sending 'SIGUSR1' to proces. Issue below command.

bash# kill -USR1 <pid of glusterfsd brick process>

> 2. How would I determine what process has that subvolume? (I only have
> one volume, I don't know what subvolume refers to in this context)

By 'subvolume' Avati meant, the brick process. Here, in your specific 
case, get the statedump of 'Brick3' from your 'gluster volume info' details.

If '$hostname:$exportdir' is your brick, you can determine which process 
is exporting that brick by below steps

bash# ssh -l root $hostname ps aux | grep $exportdir

Hope this is informative, also I hope you still have the all the process 
running (if you had done 'gluster volume stop' any time inbetween or 
restarted the node/brick process, you may want to re-create the issue 
and then take statedump.


More information about the Gluster-users mailing list