I second that. I have seen that behaviour also. It will be good to
know why this might be happening? This will cause huge impact in
production.
On Fri, Mar 4, 2011 at 1:32 AM, Netrom Nacoma <netroms at gmail.com>
wrote:> Hi,
>
> we have a large glusterfs setup with around 10 servers.
> We have three clients connecting to the servers with the fuse client.
>
> Recently one of the servers had to be taken down for maintenance each
> server is mirrored with another.
> We now have some problems with the clients hanging infinitely, on
> which client server seems to be random.
> When the client hangs the glusterfs mount must be remounted to work
> properly again.
> To test if the mount/client hangs I do a simple stat on a file.
>
> In the client log for the hanging mount/server I see a lot of this
> when the problem starts:
>
> 2011-03-04 08:14:39] E [client-protocol.c:313:call_bail]
"servername":
> bailing out frame FINODELK(31) frame sent = 2011-03-04 07:44:33.
> frame-timeout = 1800
>
> GlusterFs version is ?glusterfs 3.0.5
>
>
> Is this a bug or "normal" behavior ?
>
>
>
> Regards,
> Netrom Nacoma
> _______________________________________________
> Gluster-users mailing list
> Gluster-users at gluster.org
> http://gluster.org/cgi-bin/mailman/listinfo/gluster-users
>