[Gluster-users] in a replicated setup, how safe is it to take a brick down

Arnold Krille arnold at arnoldarts.de
Fri Apr 20 16:50:02 UTC 2012

On Friday 20 April 2012 10:28:35 lejeczek wrote:
> is it safe at all,
> a) a disaster, imagine CPU blows up
> b) deliberately, in a control fashion

My first tests show that replicated volumes are still accessible when one node 
is shut down hard or slow. And the changes are replicated back when the failed 
node reappears. If you mount it via glusterfs, the hostname is only used 
during mount to fetch the meta-data. With nfs-mount the mount only runs on 
when the mounted node runs. Of course trying to use a shut-down node doesn't 
work on either nfs nor glusterfs...

-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 198 bytes
Desc: This is a digitally signed message part.
URL: <http://supercolony.gluster.org/pipermail/gluster-users/attachments/20120420/dd786506/attachment.sig>

More information about the Gluster-users mailing list