[Gluster-users] Question about the behavior of Gluster regarding fencing a failed node

[Lalatendu Mohanty]

On 11/19/2013 09:12 AM, Xiao Bin XB Zhang wrote:
>
> Hey,
>
> I am a new babies of Gluster, and found it is very useful, and many 
> production usage.
>
> When setup our Gluster for some customer engagement project, a problem 
> rising:
>
> Suppose that I setup my Gluster with Strip + Replica 2 on my 4 
> physical nodes,
>
> so I got replica setup with 2 nodes each.
>
> At this moment, one of node failed, thus cause replica as 2 broken,
>
> at this situation, what's the right behavior of Gluster? will it just 
> use 1 remaining replica node
> to serve the service? or will it try to find another node to re-setup 
> 2 replica nodes?
>

Gluster will/should use the healthy node from the replica pair to serve 
the requests. At least one node should be healthy in the replica pair to 
keep gluster volume online.
>
>
> /Xiaobin Zhang(???)/
> /Cloud Solutions and OpenStack Development/
> /IBM China System and Technology Lab, Beijing/
> /E-mail: xbinzh at cn.ibm.com/
> /Mobile : 86-13910156791/
> /Address: Building 28(Ring Building), ZhongGuanCun Software Park, /
> /         No.8 Dong Bei Wang West Road, Haidian District Beijing 
> P.R.China 100193/
>
>
> _______________________________________________
> 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/20131120/f08c6d7a/attachment.html>