On Tue, 30 Jun 2009, Atom Powers wrote: > Given that high availability is one of my goals, what kind of storage > is appropriate? > NFS? > iSCSI?
NFS works, but leaves you with a single point of failure in your NFS server. > A GlusterFS cluster? > A storage appliance? I have used GlusterFS, it seems to be painfully slow. It is possible to do client side replication for availability, but that tends to require changing all of the client configurations whenever you want to add storage. This makes maintenance windows a bit of a headache. > It is apparent that some kind of shared storage is necessary to > migrate a VM from one compute server to another. Currently my shared > storage is an NFS server, but I'm skeptical of how well that would > perform with VMs running on it. Is iSCSI appropriate for shared > storage? Knowing that I could be running dozens of VMs, some for > critical services, high availability is a requirement. I would be very > unhappy if the storage device all the VMs were on decided to take a > vacation during my vacation. iSCSI may still wind up being a single point of failure. Other than that, to use it, you will need to either dedicate a LUN to each virtual machine so that only one host is accessing it at a time or go to something like GFS/GFS2. > What kind of storage do you use for your VMs? Still local to the systems. I need something better, but don't have the equipment, time or budget. -- Matt It's not what I know that counts. It's what I can remember in time to use. _______________________________________________ Discuss mailing list [email protected] http://lopsa.org/cgi-bin/mailman/listinfo/discuss This list provided by the League of Professional System Administrators http://lopsa.org/
