Gluster users - Feb 2019 - Client failover question

Hello,

What do you mean by a gluster 'client server' ?
A node is either a server (running the glusterd and glusterfsd/brick processes),
and you'd need at least 2 (3 is better, and the volume to be set as a
replica-3) or a gluster client (uses the gluster-client to mount a volume from a
gluster-cluster).
You could of-course also run the 'client' on the actual gluster-server
to mount to itself (wouldn't recommend that).

Nico

----- Oorspronkelijk bericht -----
Van: "Jim Laib" <jlaib01 at gmail.com>
Aan: "gluster-users" <gluster-users at gluster.org>
Verzonden: Vrijdag 8 februari 2019 17:04:45
Onderwerp: [Gluster-users] Client failover question

I may have missed this in the documentation, but is there a way to make a
Gluster client server redundant? I have several bricks setup being controlled by
one cluster client server and if that server blows, there's quite a bit of
rebuilding to do. So looking for a way to ovoid that. Any suggestions would be
most appreciated.

Thanks in advance. 

_______________________________________________ 
Gluster-users mailing list 
Gluster-users at gluster.org 
https://lists.gluster.org/mailman/listinfo/gluster-users

I'm using a single server (gluster client)  to mount a gluster replicated
gluster cvolume using two servers. so the single server is acting as the
head to the gluster volume. My users access the cluster client head using
Samba. What I'm wondering is. what's the best way to insure continuity
if
my single head server takes a nap?
And yes, I agree 3 is better and I plan to move down that path when I
reconfigure my system.

On Fri, Feb 8, 2019 at 12:51 PM Nico van Royen <nico at van-royen.nl>
wrote:
> Hello,
>
> What do you mean by a gluster 'client server' ?
> A node is either a server (running the glusterd and glusterfsd/brick
> processes), and you'd need at least 2 (3 is better, and the volume to
be
> set as a replica-3) or a gluster client (uses the gluster-client to mount a
> volume from a gluster-cluster).
> You could of-course also run the 'client' on the actual
gluster-server to
> mount to itself (wouldn't recommend that).
>
> Nico
>
> ----- Oorspronkelijk bericht -----
> Van: "Jim Laib" <jlaib01 at gmail.com>
> Aan: "gluster-users" <gluster-users at gluster.org>
> Verzonden: Vrijdag 8 februari 2019 17:04:45
> Onderwerp: [Gluster-users] Client failover question
>
> I may have missed this in the documentation, but is there a way to make a
> Gluster client server redundant? I have several bricks setup being
> controlled by one cluster client server and if that server blows,
there's
> quite a bit of rebuilding to do. So looking for a way to ovoid that. Any
> suggestions would be most appreciated.
>
> Thanks in advance.
>
> _______________________________________________
> Gluster-users mailing list
> Gluster-users at gluster.org
> https://lists.gluster.org/mailman/listinfo/gluster-users
>
> _______________________________________________
> Gluster-users mailing list
> Gluster-users at gluster.org
> https://lists.gluster.org/mailman/listinfo/gluster-users
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.gluster.org/pipermail/gluster-users/attachments/20190209/fd07aa40/attachment.html>