So in other words, for Gluster, I should configure the storage as 2 individual LUNs, and then create Replicated Volumes for High Availability. That would mean sparing 50% of the space for replicated files, more like a mirror, right? Also, how fast is Gluster at Listing files in large directories, say with 2000 - 4000 files, each? The Technical FAQ answers this question, but doesn't provide any numbers. Regards, Indivar Nair On Mon, Sep 12, 2011 at 12:39 AM, Mohit Anchlia <mohitanchlia at gmail.com>wrote: > I don't think gluster recognizes storage failures. However, you can > have your own heartbeat that would stop gluster process in case there > are storage array or frame falures. > > On Wed, Sep 7, 2011 at 9:07 AM, Indivar Nair <indivar.nair at techterra.in> > wrote: > > Hi ..., > > > > How does one deploy Gluster on a clustered storage i.e. two server nodes > > with a common storage array? > > > > Say one has, two server nodes connected to a common storage array, that > > exports 2 LUNs visible to both nodes. > > > > [ Server 1 ]------[ Storage ]------[ Server 2 ] > > > > Each server node mount a single LUN, but in case one of the node fails, > the > > other node takes over the LUN. > > > > [ Server 1 ]<------[ LUN 1 | LUN 2 ]------>[ Server 2 ] > > > > Failover: [ LUN 1 | LUN 2 ]======>[ Server 2 ] > > > > This take over is usually done using tools like 'Heartbeat'. > > > > The reason I ask is, I already have such a configuration, and would like > to > > deploy Gluster on it. > > > > Regards, > > > > > > > > Indivar Nair > > > > > > _______________________________________________ > > Gluster-users mailing list > > Gluster-users at gluster.org > > http://gluster.org/cgi-bin/mailman/listinfo/gluster-users > > > > > -------------- next part -------------- An HTML attachment was scrubbed... URL: <http://gluster.org/pipermail/gluster-users/attachments/20110912/55d30681/attachment.htm>
