VMware: virtualized SMBs do it better A new VMware survey says that small biz IT departments that embrace virtualization are more competent than those who don't. Not that you'd expect anything different from VMware. The server virtualization juggernaut needs to know what small and medium businesses are thinking - and not just the ones that buy its ESX Server …
Virtualization may be useful for a lot of things, but disaster recovery ain't one of them. Sure it can do "mild inconvenience" recovery (i.e. high-level software crash recovery), but come a water pipe rupture -or any hardware or low-level software problem, for that matter- and you're toast.
With the right approach, DR is much simpler with a VM than with a physical machine. As long as you have an appropriate host (ESX or whatever) available, you recover the VM image and go, rather than having to work with some vendor's notion of "bare-metal recovery" (possibly onto incompatible hardware) or having to reinstall the OS and then run the recovery and hope that it works. Furthermore, products like VMware's Site Recovery Manager and Platespin Protect facilitate automating the backup/recovery process, allowing a push-button recovery of your whole environment.
Other than that, you're spot on.
"Virtualization may be useful for a lot of things, but disaster recovery ain't one of them."
We have virtualised all our systems with the exception of MS SQL, which is a doddle to restore if you back it up properly. We did look at VMware ESX which works great until you try and back it up without a SAN.
We saved our pennies on ESX and then spent them on Doubletake to provide continuous replication of two virtualisation hosts running VMware server on Windows 2008 x64 to a third fileserver (VMware server will be migrated to Hyper-v at some point). We now have realtime crash consistent failover and at night we simply pause the replication and then archive the whole lot to LTO3 tape.
Our most recent offsite DR test was carried out in March. The limiting factor for the restoration of our systems was the speed of our DR providers tape drive. Everything including our AD controllers and Exchange system came back with no drama.
Virtualization can be a big help with DR.
If you can tolerate a little out of hours downtime, it's easy to keep up to date copies of virtual disks, which can be trivially restored to any server that runs your hypervisor of choice. That gives you considerable hardware independence, as well as an easier restore process.
If you can't tolerate downtime, VMs on shared storage can be snapshotted and copied. Not as easy, but you still get a much easier and faster recovery process than running on the metal.
Also, if you've virtualized and have a disaster wiping out your server estate, you've got far fewer machines to replace, saving time acquiring/commisioning them.,
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
