On Wed, 28 Oct 2009 11:24:53 -0500 Paras pradhan <pradhanparas@xxxxxxxxx> wrote: > From a last couple of weeks we are having some network maintenance > and upgrades in our data center. My xen cluster is in it. I have 3 > nodes with DRAC fencing. So when ever there is a network problem , my > cluster nodes try to fence each other and after the reboot of the > nodes, I can see a mess ( no shared storage, no vms running etc etc). > This is a test cluster so till now this is not a issue, but when I > will have a production cluster, is there any way to handle the > situation? I know if we do not have network problem i will not see > this issue, but sometimes there are upgrade, maintenance etc etc. > > Any thoughts? It's obvious that you have (at least) one SPOF in your setup. You must eliminate all SPOF's if you want true HA solution. Buy two independent low cost switches, trunk them, and connect all your cluster nodes to them. Set up LACP across switches, bond ethernet interfaces on your nodes, and finally connect your network to one of the switches. Then, when your network crew upgrades your equipment, your cluster will remain operating. -- | Jakov Sosic | ICQ: 28410271 | PGP: 0x965CAE2D | ================================================================= | start fighting cancer -> http://www.worldcommunitygrid.org/ | -- Linux-cluster mailing list Linux-cluster@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/linux-cluster