[Gluster-users] hanging of mounted filesystem (3.3.1)

[Anand Avati]

That's probably because in a 2-node replicated environment, there is no
d_off transformation (at the DHT level) and therefore the bug is "avoided".

Avati

On Thu, Jan 31, 2013 at 5:58 PM, Michael Colonno <mcolonno at stanford.edu>wrote:

>         Doing more searching, it appears there are several posts on various
> forums describing the same issue with Gluster deployments of 4+ nodes (e.g.
>
> http://serverfault.com/questions/453345/certain-operations-like-ls-hang-in-4
> -node-gluster-setup). None of these posts are answered which doesn't fill
> me
> with confidence. Can anyone confirm a 4+ node cluster with replication that
> does not show this (indefinite hang on ls or ll) behavior?
>
>         Thanks,
>         ~Mike C.
>
> -----Original Message-----
> From: gluster-users-bounces at gluster.org
> [mailto:gluster-users-bounces at gluster.org] On Behalf Of Michael Colonno
> Sent: Thursday, January 31, 2013 5:48 PM
> To: gluster-users at gluster.org
> Subject: Re: [Gluster-users] hanging of mounted filesystem (3.3.1)
>
>         I rebuilt the volume with only TCP transport; same issues occurring
> so RDMA must not be the issue here. With a glusterfs mount (which is
> successful) I can copy files into the mounted volume but any command to
> list
> the contents (ls or ll) hangs forever while ls (for example uses ~ 5% of
> the
> system's CPU and glusterfs uses ~ 50% of the system's CPU). I've never seen
> this behavior before with my other Gluster deployments, even much larger /
> more complicated ones. This one is pretty ordinary. Any advice appreciated.
>
>         Thanks,
>         ~Mike C.
>
> -----Original Message-----
> From: gluster-users-bounces at gluster.org
> [mailto:gluster-users-bounces at gluster.org] On Behalf Of Michael Colonno
> Sent: Thursday, January 31, 2013 4:50 PM
> To: gluster-users at gluster.org
> Subject: Re: [Gluster-users] hanging of mounted filesystem (3.3.1)
>
>         Tried the proto=tcp setting, same error unfortunately...
>
>         Thanks,
>         ~Mike C.
>
> -----Original Message-----
> From: Robert Hajime Lanning [mailto:lanning at lanning.cc]
> Sent: Thursday, January 31, 2013 4:28 PM
> To: Michael Colonno
> Cc: gluster-users at gluster.org
> Subject: Re: [Gluster-users] hanging of mounted filesystem (3.3.1)
>
> On 01/31/13 15:54, Michael Colonno wrote:
> > Updating this thread: I tried to switch over to NFS mounting. When I
> > try to mount the volume with the following line in fstab:
> >
> > node1:/Volume /tmp/mnt nfs defaults,_netdev,vers=3 0 0
>
> node1:/Volume /tmp/mnt nfs defaults,_netdev,vers=3,proto=tcp 0 0
>
> --
> Mr. Flibble
> King of the Potato People
>
> _______________________________________________
> Gluster-users mailing list
> Gluster-users at gluster.org
> http://supercolony.gluster.org/mailman/listinfo/gluster-users
>
> _______________________________________________
> Gluster-users mailing list
> Gluster-users at gluster.org
> http://supercolony.gluster.org/mailman/listinfo/gluster-users
>
> _______________________________________________
> Gluster-users mailing list
> Gluster-users at gluster.org
> http://supercolony.gluster.org/mailman/listinfo/gluster-users
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://supercolony.gluster.org/pipermail/gluster-users/attachments/20130131/97788b1f/attachment.html>