Server Fault is a question and answer site for system and network administrators. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

We recently got an inline generator installed at our facility that backs up our IT room. It detects the outage and kicks on within about 20 seconds of an outage. For those 20 seconds between the time of the outage and when the generator takes the load over, the equipment is running on battery power supplied by our current line-interactive UPSes. Unfortunately, the power delivered by the generator has a large enough variance in voltage that the UPSes alternate between battery and generator power several times per minute. I know this can be mitigated by double-conversion / online UPS systems, but I'm wondering if the current setup will suffice (aside from the screeching of the current UPSes).

My thought is that the only way to tell will be to run off of the generator for an extended period of time and monitor to see if the UPSes are running on generator power enough of the time that they can keep their batteries charged up. There's no doubt that the screeching will be annoying to anyone in the vicinity, but employee workstations will have crapped out within 5 minutes of any real outage anyway so it's not as though they'll have much work to get done. Our basic need is simply to keep servers, networking equipment, and phones up and running indefinitely in an outage.

share|improve this question
up vote 1 down vote accepted

Probably the simplest approach would be to monitor battery charge levels in the UPSes during a periodic test (you do run periodic generator tests, right?) and see if they're maintaining, losing, or gaining charge levels over time.

share|improve this answer
We do run generator tests. They're scheduled to run once a week without switching over. We'll schedule something a bit less frequently to make sure that it's properly kicking over to generator during outages. I suppose I'll just have to monitor the battery levels during one of those load tests. – pk. Apr 28 '11 at 2:45
The only other approach I could imagine is to measure amp draw incoming to the UPS and amp draw outgoing, then determine how much time is spent online versus off and multiply accordingly. This would be strictly back-of-the-envelope territory, though; you'd have to account for all sorts of losses that you wouldn't necessarily see in the numbers, but if you're covered by a factor of two or three (or short to a similar degree), that would emerge pretty quickly. – BMDan Apr 28 '11 at 18:15
I guess my main question is whether or not this is unreasonable to expect UPSes to be able to handle the constant switching on and off battery power during an outage. Let's say we're down for 8 hours and they switch on/off 3 times per minute. That'd be 1440 failovers for the 8 hour duration. Is that too abusive to the UPSes or is this OK for extenuating circumstances such as an outage? – pk. Apr 28 '11 at 21:51

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.