Uninterruptible Power Supplies are extremely rugged beasts -- which is both good news and bad news for storage administrators who rely on them. The good news is that UPS will continue to work even under conditions of partial failure, which means your critical
For that reason, you shouldn't rely on indicator lights and other passive signs to monitor the condition of your UPS. UPSs should be tested regularly.
In the case of a small UPS supplying a single server, this kind of testing is fairly simple. It involves cutting power to the system to see if the UPS picks up the load and carries it for the specified period, while your system automatically shuts down and issues the appropriate alarms. You should also test battery capacity to see how long the UPS will carry your load, and check the line voltages as the UPS takes over to make sure there are no serious spikes and surges. Also, don't forget to check the condition of the batteries regularly and replace them according to the manufacturer's recommended schedule.
With a large UPS supplying an entire data center, the testing is more complex. It can also be more dangerous because of the currents involved. If your people don't have the proper training, it might be worthwhile to outsource the job.
A test of a large UPS should include calibrating the unit, checking all the protection settings, checking and tightening connections and inspecting capacitors for physical signs of failure. In addition to the performance tests of switchover, transient response, battery life and battery condition, a complete test of a data center-sized UPS should include harmonic analysis at various loads, a check of filter integrity and a test of response to module failures in the UPS.
Obviously, this kind of testing is expensive, whether you do it yourself or hire an outside engineering group to do it for you. Even testing a small UPS disrupts the normal operation of the equipment the unit protects. You should consult with your UPS supplier to determine what kinds of tests are recommended and how often they should be performed.
For more information:
Backup school: Lesson six, Protection from hardware loss
About the author: Rick Cook has been writing about mass storage since the days when the term meant an 80 K floppy disk. The computers he learned on used ferrite cores and magnetic drums. For the last 20 years he has been a freelance writer specializing in storage and other computer issues.
This was first published in September 2004