[SOLVED] Why did parity start?

[nraygun]

I have my system setup to do the parity check on the first of the month.

For some reason, it started a parity check this morning at about 5am. The uptime of my system seems to indicate it may have rebooted by I can't tell why.

I looked through the logs and see the entry below, but I can't tell what lead up to this.

Is there a copy of the logs from the last time it was up and running that might show, say, a power outage?

Anywhere else I should be looking?

Nov 4 04:59:30 flores emhttpd: unclean shutdown detected

 

flores-diagnostics-20201104-1709.zip

Edited November 6, 2020 by nraygun

[Squid]

1 hour ago, nraygun said:

Is there a copy of the logs from the last time it was up and running that might show, say, a power outage?

No.  At least not if you didn't previously have the syslog mirroring to the flash drive (only recommended to do for constant / recurring reboots)

[Frank1940]

Does your server have a reset button on the case.  Is there a possibility that someone or an animal (cats are a prime suspect here...) might have pushed it?   IF there is such a button, you might want to disconnect the switch leads from the MB.

 

That have been some a couple of instances where a defective PSU caused a server to reboot.

 

Powerages normally cause a complete shutdown requiring manual action on your part.  (There are a few BIOS's with a setting that will restart a server on power restoration but I, personally, would not recommend using that setting to do so...)  Do you  have any other signs of a powerage?

[nraygun]

No other signs of power outage at that time. The server is in the basement (Dell R710) and our cat doesn't have access to it.

I have a UPS on the server and I think I have BIOS to start back up when power is restored.

I was hoping a log was written before the shutdown process to show why the shutdown/restart occurred.

Any other ideas?

Edited November 5, 2020 by nraygun

[Frank1940]

If it was a "triggered" reboot that would  have been caught in the normal logging process, you would have had a clean shutdown.  There is the Syslog Server that you can setup for these types of problems.   You can read about it here:

 

      https://forums.unraid.net/topic/46802-faq-for-unraid-v6/page/2/?tab=comments#comment-781601

 

You might want to verify that the UPS is functioning correctly.  To do this, shut the server down.  Now connected enough incandescent light bulbs to it to load it to at least 50% of the UPS rating.  Now see if it will power this load for five to ten minutes.  (I have seen UPS batteries that don't last ten seconds!)

 

I would be checking to see if the server is set to restart on power restoration.  IF it is, turn it off until you have solved this problem.  You need to know if the server is rebooting or shutting down.  (BTW, I consider this option to be a bad practice.  While it may seem to be an acceptable practice, it actually wrought with problems is the primary power is unstable after restoration because of utility/natural-disaster problems.  Most UPS's will not support a second power outage without eight to twenty-four hours to recharge the batteries!) 

[nraygun]

Yikes!
I think I found the problem.

It started another parity check at around noon right after it finished the one from this morning. And it only took 8 hours which is way too short.

Then the server went offline altogether.

So I went down into the basement and found that the system LEDS were all blinking(power, the HDDs, etc.) and the LED on the power supply was flashing between green and amber. I removed the cable and reseated the power supply. Still the same blinking lights.

I only run a single power supply on my R710 to save power so I still had the second power supply. I popped that in and it started up immediately.

Of course now it's running a parity check but this is to be expected since the rug was pulled out from under it.

I'll keep an eye on it as it runs through this latest parity check but it seems like this may have been a power supply issue.

 

[nraygun]

Just to close this out (I think), the parity is still running after ~19.5 hours and hasn't restarted. My backup script (borg based) is taking a little longer than normal but I'm guessing it's because it was interrupted mid-task.

Concluding it was the power supply. I'll get on eBay and buy a couple of backups and consider running two PSU's in redundancy mode. It'll eat up an extra 5W or so, but it might be worth it.

Thanks to those who responded with ideas!