As you probably know by now if you’ve ever read this blog before, I like to ask questions of my readers. It’s a way for me to gain some added industry experience without actually having to live through
any some of the common pitfalls of the day-to-day life of a Systems Administrator, and typically you guys are pretty helpful and informative while being only slightly snotty…
My question today relates to power. You know, that sparky stuff that comes out of those holes in the wall and makes things go (and hurts you if you stick a key in one of those funny holes… go figure).
Recently, I’ve finally gotten fed up with Duke Power’s service. Since there’s no real alternative or forseeable change in the near future, I decided that it was time to put a stop to the random power flashes and guaranteed repetative flashes during thunder storms. I’ve got small consumer (ie: one-PC) UPSes here at home, but they’ve all died along the way (it’s not just the battery, something’s actually happened to the rest of the UPS itself. Power works on the surge-protected outlets, but not on any of the UPS-backed ones).
I was looking at pricing out UPS systems to replace my dead ones the other day, and since I have no idea about electrical stuff aside from the fact that you never NEVER NEVER want to cross those two wires (oh, and that it’s always the green wire you cut… or was that red…?), I went to APC’s website to use the online configurator they have for selecting an appropriate UPS system.
Well, it looks like my needs may be a tad more than the average user’s household needs. They quoted me at a standard wiring-closet (1 to 3 racks) UPS system that runs about $5,000.
Now, since that’s clearly out of the question, I’ve started formulating other plans that have yet to be set in motion. Still, this got me thinking. Most UPS systems are designed to support a certain amount of equipment for under 10 minutes (on average 6 to 7 minutes is common). At work we have a series of rack-mounted equipment that will support about a dozen boxes each, with the added battery packs. Still, the runtime isn’t going to be stellar (we’re probably looking at 15 minutes, tops).
So my question is: what’s common in the industry? We’re not talking data-center stuff here. We’re talking a normal 8 to 5 shop (give or take) that’s open Monday thru Friday and closed all major holidays anyway. What kind of setup do you have at work? Do you even bother? Is the idea really just to survive the odd power flash without data loss (since the desktops on the floor that are unprotected aren’t going to be up through it, and operations won’t be immediately resuming), or have you geared your setup for a more 24x7 approach (and why)?
In reality, at any point in time during the day, we could probably stand to lose power without warning and still not lose any data. Sure I’d spend an hour or two bringing systems back up, making sure they came up cleanly, and that everything was running smoothly again, but is that the only reason we’ll drop $10,000 for a UPS that’s going to keep us going only for a handfull of minutes?