On Mon, Dec 5, 2011 at 08:27, Ben Scott <[email protected]> wrote: > On Mon, Dec 5, 2011 at 7:56 AM, Andrew S. Baker <[email protected]> wrote: >> He's going to pay more per GB for local storage than for SAN storage >> (for the SAN devices being recommended), and he'll have more >> flexibility for how that space is carved up for VM usage, backups, etc > > Backups are going to external disks which will be rotated off site. > Basically taking our existing backup plan and s/tape/disk/ with it. > > SAN snapshots are nice but are not backups, as Microsoft discovered > when the Sidekick SAN went tango uniform. > > So I'm just carving up the space for the various VMs. I don't see > how a SAN helps there. All it does is add significant complexity, > which means it adds failure cases. "A twin engine airplane has twice > as many engine problems as a single engine airplane." Redundancy is a > very good reliability technique, but one must take care that the cost > (including non-monetary costs, like complexity) doesn't exceed the > benefits. > > If the situation changes and a SAN starts to make sense, migrating > VM storage off DAS to the SAN at that point is easy, for some value of > "easy". > >> I find it highly beneficial to have two VM hosts. > > I'm not arguing it's not beneficial. I'm saying for the current > situation it's not cost effective. It would double my server cost, > *plus* add the cost of the SAN, which for an environment this small is > non-trivial. That would eat up a ton of budget that I could spend > elsewhere on things that need it more. > > Meanwhile, I can get a 1-hour-typical parts response for 20% the > cost of the single server equipment. > > Software failure on the VM host is the remaining scare spot. But > heck, reinstalling the entire OS from scratch doesn't take *that* > long, and the nice thing about a Server Core+Hyper-V host is there > isn't much to configure. > > If I had a huge pile of cash just waiting to be spent, sure, I'd go > for the multiple VM hosts and SAN. > > -- Ben
For an installation of your size, I mostly agree. As I've stated before, there are two disparate approaches: o- Don't keep all your eggs in one basket o- Keep all your eggs in one basket and watch it very carefully. I think you're on the right track. However, part of watching it carefully is having vendor hardware support that covers you sufficiently, and you're on top of that as well. If you had more VMs and more cash, I'd actually recommend three hosts instead of two. That way, if one of them fails, the other two can more easily pick up the load and still function close to normally. Kurt ~ Finally, powerful endpoint security that ISN'T a resource hog! ~ ~ <http://www.sunbeltsoftware.com/Business/VIPRE-Enterprise/> ~ --- To manage subscriptions click here: http://lyris.sunbelt-software.com/read/my_forums/ or send an email to [email protected] with the body: unsubscribe ntsysadmin
