Periodic and brief "Insufficient Runtime Available" messages

[Anonymous user]

One of my clients has an odd, recurring problem. They have two identical SU3000RM2U UPS units. One, "UPS B", works fine and gives no trouble. The other, "UPS A", often logs an "Insufficient Runtime Available" message, which is then followed by a "Sufficient Runtime Available" message about nine seconds later, on average.

UPS B (the good one) has a load of 50% and a runtime of 13 minutes, according to PowerChute.

UPS A, on the other hand, has a load of 54% and a runtime of only 8 minutes. The battery was just replaced on May 2, and a manual calibration was run on May 3. The problem predates the new battery.

Both UPS units are a bit over two and a half years old. My client is running PowerChute Business Edition Deluxe 8.01.

The insufficient runtime messages most commonly happen around the 7:00 to 8:00 PM time window, but they've happened at numerous other times as well. There is no obvious date pattern; it can happen multiple times in one day, or it can go 3-4 days or more without happening. I've tried cross-referencing these events with the data log, but I see no discernible pattern or cause.

Each UPS is plugged into a different circuit, and that may account for voltage drops on only one UPS, but if that's what it is, such drops aren't visible in the log. And that still wouldn't explain the insufficient runtime message.

I've tried setting the shutdown sequence for as little as four minutes, but the problem still occurs.

Any ideas?

[Anonymous user]

What sort of loads are connected to the UPS that is reporting these messages? Is it possible that a device powered by this UPS has a temporary demand for more power which is triggering the message? Is there something that commonly happens in the 7 to 8 PM window?

As an example, I have a file server connected to an AP7900 swiitch (which monitors the current draw of connected devices), and if I reboot the server (which causes 16 disk drives to spin down and back up), it will report "A near overload condition exists" followed by "A near overload condition no longer exists".

If you have something like an automated tape library that performs a backup around 7 PM, the added power consumption of the library mechanism moving tapes around might cause the UPS to re-evaluate its runtime and report the messages you're seeing.

Would it be possible to swap loads between the UPS systems (plug the "A" loads into "B", and vice versa)? If the problem moves to the other UPS, then it is due to one of the loads. If it stays with that UPS, swap the input circuits (plug "B" into the outlet that provides power to "A", and vice versa). If the problem moves, then it is an issue with the input circuit, while if if it stays on the same UPS, then I'd say there is some issue with that particular UPS.

I'm more familiar with the web/SNMP cards, as I don't use PowerChute here. If PowerChute works the same way, what you want is something like "Alarm if runtime under X minutes", not "shutdown delay".

[Anonymous user]

No robotic tape libraries or anything like that. The main things hooked up to these UPS units are servers. I've cross referenced the days and times that the messages occur in a spreadsheet and graph, and the results, while interesting, only deepen the mystery. All of the things I can think of that might cause an increased power draw (backups, maintenance scripts, updates, etc.) typically happen in the late night/early morning hours, when it appears we never get an alert. And again, the alerts are sporadic and brief. I can't predict what day they will happen, but when they do, I can predict the time with a reasonable degree of accuracy. Typically it's around 7-8 PM, but never at exactly the same time. Less often, it can happen somewhere around 5-7 AM or 11 AM - 2:30 PM. It also appears that the less common times never happen on Friday or Saturday, but the 7-8 PM occurrences can be on any day.

I can do a circuit or load swap the next time I do server maintenance on a weekend. Either is easy enough to do during down time. I'll probably try the load swap first.

In PowerChute, I see no option for an alarm/run time setting. The only place I see to specify a number of minutes in relation to run time is "When power fails begin a shutdown procedure: After UPS has been on battery for minute(s)". However, given its location, this setting appears to relate to the automated shutdown of servers, a feature my client isn't using. I see nowhere else to change any run time related settings.

[Anonymous user · Accepted solution]

You're correct. I only see the "Alarm if runtime under X" in the larger units, like my Symmetras.

Hopefully someone from APC can comment on what triggers the "Insufficient runtime available" message you're seeing.

Sorry for the incorrect info. It will still be useful to swap loads to see if the problem moves or not.