Hello all, I've been running my unRAID server for 4 years or so now with basically no issues. A few months ago, I noticed that it had seemingly randomly shut down (unclean) and was doing an unscheduled parity check. This has since happened 5 or 6 times with increasing frequency. I am running it with an APC UPS but I also know the power has not been turning off at the house during these times. The most recent times it has happened anywhere from 5-7 days apart. I just checked my bios and the "turn back on after power interruption" feature is turned OFF. I'm unsure why/how the server is turning itself back on after it goes down. It is at least going down according to unRAID, since the uptime counter always resets at these occurrences.

 

Recently, I installed Fix Common Problems and ran Troubleshooting mode. I have attached those files but I am unable to see anything obvious in them but maybe you'll have more luck. I also ran Memtest86 overnight which did produce one error. Through my research it seems this error is unlikely to cause issues, but let me know if that is inaccurate. It failed 1/4 of test #13, the Hammer test, resulting in one error.

 

Nothing has been recently changed hardware wise with the server to coincide with these issues. As far as dockers, I'm just running Sab, Transmission, Couchpotato, Sonarr, and Roon. I began running Roon not too long ago but I do not believe it coincides with these issues starting, but I could be wrong.

 

My hardware is:

ASRock FM2A85X-ITX

AMD A6-5400K

PC Power and Cooling Silencer MKIII 400 watt PS - PPCMK3S400

Kingston 4GB RAM x 2 - KHX1600C9D3

A number of 8tb and 3tb HD's.

 

I am concerned about these random shutdowns because I obviously don't want to end up losing any data. This current build is getting a little old/underpowered so if the best course of action is to just toss the mobo/processor/RAM/power supply and upgrade I'm willing to do that. If it's just a minor issue that is easily fixed I'd go with that option if possible.

 

Any thoughts on what may be causing this issue? Thanks in advance, I appreciate it!

syslog.txt

FCPsyslog_tail.txt

MemTest86-Report-20170928-042059.html

Edited September 28, 20178 yr by highgear

12 minutes ago, highgear said:

I also ran Memtest86 overnight which did produce one error.

 

This already indicate a critical problem of your hardware. From past O.C. experience, Memtest86 reallly not a useful strees test tools, it can't quickly detect problem or just better than nothing.

 

Suggest do a visual check any capacitor exploded on mainboard and try only install one memory stick for test first.

Thanks for the response. I'll do a visual inspection and report back. So are you suggesting running Memtest86 with just one stick or what?

As mention Memtest86 just better than nothing, but seems no other option, so pls test one stick.

 

I won't run strees test under unRAID, I will boot in Windows ( from standalone harddisk ) for test and make sure system were stable.

Edited September 28, 20178 yr by Benson

Ok I visually inspected the mobo and I'm not seeing any issues. I'm now running a memtest86 cycle with just one stick of RAM. Just so you know I ran another set of memtest86 tests with both sticks in earlier and no errors came up.

Thanks, pls share the result.

 

3 hours ago, highgear said:

The most recent times it has happened anywhere from 5-7 days apart.

 

But can't conclude in short time.

Edited September 28, 20178 yr by Benson

One possible cause of this type of problem is intermittent failures with the power supply.  I am not saying that it is but keep this in your mind as you move forward.  It is not an expensive component to purchase and one of the easier ones to change out. 

 

Most experts also feel that a failure in Memtst is a cause for concern...

22 minutes ago, Benson said:

Thanks, pls share the result.

 

 

But can't conclude in short time.

Will do.

 

8 minutes ago, Frank1940 said:

One possible cause of this type of problem is intermittent failures with the power supply.  I am not saying that it is but keep this in your mind as you move forward.  It is not an expensive component to purchase and one of the easier ones to change out. 

 

Most experts also feel that a failure in Memtst is a cause for concern...

Yeah it seems like PS or RAM are the most common causes for this type of problem. I still don't understand how/why the system would boot back up if it were either of these issues instead of just turning off. So even with the one memtest error you still think it could be the PS? RAM would certainly be the easiest thing to replace in this scenario...

The one of the PS voltage output voltages may drop low enough to trigger a reboot or the reboot/power-on/off circuitry in the PS is wonky and signals for a PS reboot.  Part of this circuitry is on the MB and it could also be defective and basically do the same thing.  (I seem to recall that someone had a case where a very sensitive on/off switch on the front panel would randomly trigger a power-off. Cats have also been known to hit the switch when playing around the case.)

 

RAM is easy to replace.  If you are going to purchasing new RAM, go to the MB manufacturer's website and see if they have some recommended modules.  Some MB's will take about any RAM module and others are really, really picky about the ones they will 'tolerate'.  Particularly, if you want a lot of RAM and are going to be using all four slots. 

Ok I'm ordering a new power supply but I'm going to hold out on buying new RAM at this point. There have been no more errors from memtest86, I ran it through 4 cycles each using one stick only in each slot.