Hi James, Well, uCarp does VRRP in a very simple way but it's not the solution for multiple nodes. I don't know keepalived. Wil it manage with a single intances the placement o IP addresses accross all nodes and take-over by any other one randomly if a node goes down ? Fernando ________________________________________ From: James [purpleidea at gmail.com] Sent: 21 August 2012 18:30 To: Fernando Frediani (Qube) Cc: 'gluster-users at gluster.org' Subject: Re: CTDB with Gluster On Tue, 2012-08-21 at 16:33 +0000, Fernando Frediani (Qube) wrote: > Just have a simple and easy IP distribution/take-over system in the case any nodes fail. So what you want is VRRP. Why not use keepalived ? It's very easy to setup. James > > Fernando > > -----Original Message----- > From: James [mailto:purpleidea at gmail.com] > Sent: 21 August 2012 15:54 > To: Fernando Frediani (Qube) > Cc: 'gluster-users at gluster.org' > Subject: Re: CTDB with Gluster > > On Tue, 2012-08-21 at 08:40 +0000, Fernando Frediani (Qube) wrote: > > Has anyone used CTDB(http://ctdb.samba.org/) for IP failover/balance > > between nodes using Gluster ? > I use keepalived at the moment, to provide a virtual ip for the gluster cluster (to use as the mount ip) but in the future I'll probably switch to cman/corosync based cluster management which will manage a virtual ip "resource". > > > I guess that?s what was used on the old commercial version of Gluster > > Storage Platform. > > > > Initially I had thoughts on uCarp, but CTDB seems a much better fit > > for this type of environments. > I don't know anything about CTDB, but my one minute viewing of their webpage makes me think otherwise. What are you trying to do ? > > > James > > > > > > > > > Does it do the job well and fast ? > > > > > > > > Thanks > > > > > > > > Fernando > > > > > > > > > > _______________________________________________ > > Gluster-users mailing list > > Gluster-users at gluster.org > > http://gluster.org/cgi-bin/mailman/listinfo/gluster-users >
