I've done quite a bit of testing on this the last few months, and high
availability configurations do work, you just need to know exactly
what's going on when designing your configuration. Look in the archives
for the past 30-45 days, and you'll see quite a lot of info about HA
configurations and how to configure them for robustness.
In short, my suggestion would be to move the unify to the client side
and see how it works for you.
Anand Babu Periasamy wrote:
---------- Forwarded message ----------
From: CityToo - Benjamin Loire <bloire@xxxxxxxxxxx>
To: gluster-devel-bounces@xxxxxxxxxx
Subject: High availability doesn't exist in version 1.3 glusterfs ?
Date: Tue, 30 Oct 2007 01:34:44 +0100
User-Agent: KMail/1.7.2
MIME-Version: 1.0
Content-Type: text/plain; charset="iso-8859-1"
Content-Transfer-Encoding: 7bit
Content-Disposition: inline
Message-Id: <200710300134.44211.bloire@xxxxxxxxxxx>
X-detected-kernel: by monty-python.gnu.org: Linux 2.6 (newer, 2)
Without using dns round robin remote host, high availability doesn't
exist in version 1.3 glusterfs ?
I test, and if I unplug eth0 of one node server, the connexion between
client and storage cluster is stopped !
Why translator unify with scheduler "rr" is not able to use
automatically second server connexion to keep high availability ?
Because if I have 100 nodes server and only one crashs, the storage
cluster becomes down !!??
Do I have wait for version 1.4 for high availability with easy
translator to configure ?
Else, how do you install easy dns round robin in local without using
Bind DNS?
Benjamin LOIRE
_______________________________________________
Gluster-devel mailing list
Gluster-devel@xxxxxxxxxx
http://lists.nongnu.org/mailman/listinfo/gluster-devel
.
--
-Kevan Benson
-A-1 Networks
