UPS issues - new server, failing shutdown

[tiwing]

hi, looking for some advice on UPS setup. 

 

I had my old server connected to a UPS for years and it always worked perfectly - controlled shutdowns were fine and it shut off the UPS as it should. (motherboard and processor from a Lenovo s20).

 

I very recently switched to a supermicro and unfortunately have had one power failure and one brownout in the past three weeks both times resulting in an unclear shutdown followed by a 24 hour parity check. This morning I pulled the plug on the UPS to see what would happen and within 5 seconds the server powered off... unclean shutdown and now it's doing another parity check.

 

Same UPS. Same cables. Same USB boot device -> same settings exactly. Different server hardware. Oh and I recently upgraded to 6.9 but something tells me that is not the culprit here.

 

Similar load. The s20 was less power hungry, but I also had two POE access points, modem, router, and a couple of switches on the UPS before, and I've moved them now to a different UPS entirely so the draw on the UPS is just the server and actually about the same. roughly 240W average on a 750W UPS. Runtime is estimated at 15 minutes normally.

 

On the unraid dashboard I can see the readout of power draw and it changes and corresponds to the UPS screen, so from that respect I can confirm the USB cable is functioning. 

 

Diagnostics look all OK on the UPS settings page.

 

Do I need to do anything to make the UPS work properly again?

 

Thanks

tiwing

 

kscs-fvm2-diagnostics-20210310-0953.zip

[PeteAron]

do you have the apcupsd running on your server?

[Squid]

1 hour ago, tiwing said:

connected to a UPS for years

Have you considered that the battery is no good?

[tiwing]

1 hour ago, kimifelipe said:

do you have the apcupsd running on your server?

I assume that if I can see battery level, load, and all the metrics through unraid that it is running. Perhaps not a good assumption?

 

37 minutes ago, Squid said:

Have you considered that the battery is no good?

I did - but the fact it worked perfectly a month or two ago on my old server and gave proper run-time and shut down led me to eliminate that as a possibility. I guess it could be so... I do have 2 UPSs so I can test that, although the other one is lower power (500W). I'll swap power cords tomorrow once the parity check is complete and try again. The beauty of redundant power supplies... don't have to shut it down to swap UPSs!! Will report back tomorrow prob around this same time.

 

 

[PeteAron]

yes, of course, you're right about that tiwing.

[ConnerVT]

It certainly sounds like a UPS hardware issue.  Or user error.

Let's take UNRAID and computers out of the picture for a moment:  You have a 250W load on the 750W UPS, and it is reporting that it should run for about 14 minutes.  You unplug the AC power, and it shuts down 5 seconds.

 

Even if it was a dumb UPS, with no communication to the server, it should last longer than 5 seconds. 

 

Two possibilities here:

1.  Battery is worn out, or component failure in the UPS;

2. You are not plugged into a backup battery outlet in the UPS.  Many units have outlets that aren't backup battery powered.  Just offer surge protection.  You're not plugged into one of those, are you?

[tiwing]

42 minutes ago, ConnerVT said:

Or user error.

That's always my first go-to. I'm not an expert in any of this stuff.

 

43 minutes ago, ConnerVT said:

Two possibilities here:

1.  Battery is worn out, or component failure in the UPS;

2. You are not plugged into a backup battery outlet in the UPS.  Many units have outlets that aren't backup battery powered.  Just offer surge protection.  You're not plugged into one of those, are you?

2) I double checked that before pulling the plug this morning... I HAVE done that before and spent an hour scratching my head going WTF... !! But not this time. The UPS does have both, but I'm plugged into the "backup" section with for both power supplies.

 

1) will test tomorrow with the other UPS... it's old too, so it won't necessarily be conclusive. But it will be an indication!

[S80_UK]

One other avenue to explore - you can also have the UPS check the batteries / runtime itself.  It's known as a re-calibration.   It is worth doing this once in a while, especially if your UPS is not doing regular self-tests.  This post shows you how...

 

 

Edited March 10, 2021 by S80_UK

[tiwing]

swapped power supplies to the other (lower powered) UPS, ensured connection in unraid, and pulled the plug. totally normal controlled shutdown and no parity check after restart.

 

So clearly something is wrong with the other UPS, even though it worked fine a month ago. Must just be coincidence...! (I don't like coincidences). In all your opinions is smart UPS worth the extra cost? UPS will only ever have one server attached to it. In my case, I start a controlled shut down after 60 seconds of no power and have timeouts set to 5 minutes. It doesn't have to keep running forever, it just has to have power to shut off and the shutdown also turns off the UPS. ONce power is restored, BIOS is set to "always on" after power loss so unraid starts back up... 

 

Cyberpower is cheaper than APC by a little and seem to have good reviews. Opinions on brand in the 1000-1500VA range?

 

cheers

[planetwilson]

On 3/10/2021 at 6:55 PM, ConnerVT said:

 

Two possibilities here:

1.  Battery is worn out, or component failure in the UPS;

2. You are not plugged into a backup battery outlet in the UPS.  Many units have outlets that aren't backup battery powered.  Just offer surge protection.  You're not plugged into one of those, are you?

 

OMG...I have been running for several years it would seem plugged into the surge only socket....I only came to this post as I just had a power cut and got an immediate parity check which didn't feel right.