I have a UPS. My power went out for a few minutes and the UPS kicked in. Power comes back on, I check the server and now I have write errors. The server never shut off as the power came back on before it hit the shutdown threshold. I've attached the diag. I've started the data rebuild. Only 2 disks were affected. 

tower-diagnostics-20250316-1917.zip

My UPS is: CyberPower 1500VA / 900 Watts True Sine Wave Uninterruptible Power Supply (UPS). 

I have a JBOD attached to the server also on the UPS. 

The error occur time exactly power failure time, it look like not only enclosure disks suffer the problem.

 

Mar 16 13:53:39 Tower apcupsd[15677]: Power failure. 

 

Pls note when power failure,  line-interactive type UPS will switching to and switching back, so there have two power interruption in very short time. If your PSU haven't enough hold-up time or UPS switching time too long, then problem will happen.

 

Edited March 17, 20251 yr by Vr2Io

16 minutes ago, Vr2Io said:

The error occur time exactly power failure time, it look like not only enclosure disks suffer the problem.

 

Mar 16 13:53:39 Tower apcupsd[15677]: Power failure. 

 

Pls note when power failure,  line-interactive type UPS will switching to and switching back, so there have two power interruption in very short time. If your PSU haven't enough hold-up time or UPS switching time too long, then problem will happen.

 

So what's the solution? Do I need to switch the UPS to a different type? You mentioned line interactive. I see there are online and double conversion UPS. Would you recommend me replace this one and get an online UPS? Or is there another solution? 

In my opinion, line-interactive type UPS are most common and practical, other type UPS are expensive and use lot of electricity.

 

So, the best way is to test power interruption periodic, make system ( and disks ) in full load  to draw maximum power ( but don't put disk in mount state ).

Edited March 17, 20251 yr by Vr2Io

Do you have any recommendations on how to do this test? I'm not sure what I should be looking for either. I can switch the power off while under full load. But I already know that when I switch over I get write errors. Are you suggesting the UPS should be replaced? 

1st, make system start in maintenance mode

 

 

2nd, Spindown all disks and take the minimum power usage.

  

 

 

3rd, Spinp all disks and take the idle power usage.

 

 

4th, Perform parity / disks check without correction

 

  

 

5th, Your could also stress the CPU by install plugin "core-freq'" and run stress test by shortkey "O" - Atomic Burn

 

 

In each power state, press UPS test button to verify which power usage level start have problem. As mention before, PSU also could be problem source. So you need test carefully and systematic.

Edited March 17, 20251 yr by Vr2Io

Is this:

Something I can add to unraid or is it available on a newer version? I'll give it a try when I'm finished rebuilding the disks. 

11 minutes ago, topmoviesonlime said:

Is this:

Something I can add to unraid or is it available on a newer version? I'll give it a try when I'm finished rebuilding the disks. 

You need buy a power measurement device, you can install a plugin to show the reading in Unraid.

 

 

But I add-on use homeassistant to capture those data and got in advance.

 

Edited March 17, 20251 yr by Vr2Io

I'm pretty sure I'm a dolt. The UPS has 6 battery protected outlets and 6 surge protected outlets. I had some of my array plugged into the surge protector portion only. Appreciate your time with this!