Server Dead after power down.

[Dela7460]

A few nights a go we had a bad storm and we lost power for about 14 hours. The Unraid server is on its own UPS and all other switches and Starlink on its own UPS too.  So when power goes down its not really a problem; Since this was a extended outage I got the email alert from unraid saying it was shutting down due to power levels being around 20% left.

 

Once power was restored the next day I went to power everything up and the server didn't do anything. I did a bunch of testing and figured it was the power supply. I remove the power supply and tested it and at first it was bad but then found out later that the power supply I have auto resets itself after a fault if left unplugged for some time. Either way it eventually powered on when I had it out side the  system. I put it back in and nothing... 

 

Not wanting to play a guessing game with this auto reset bull I order a new power supply just for the heck of it. It came in today;  I tested it on the bench and it turned without issue. Popped it into the server and still nothing... Not a flash or anything. I tried unplugging everything and  all the drives and still nothing. Also removed all ram and even tried one stick at a time.

 

I tried moving it off the UPS and tried other plugs and still nothing. It was a bad lightning storm but I don't think we go hit. All the other systems around the house are fine even another server on the same circuit but different UPS. Even those that are not on a surge protector.

 

Guess I'll tear the whole thing down and rebuild it. Something is causing a short somewhere; just very odd that it happened after a gracefully shutdown. I pulled the Unraid server USB drive and tested it; its fine, logs show everything good. 

 

Just bad luck I guess.

 

Wondering if anyone else has ran into a issue like this. Server lives in a sever closet; nice and cool, no one goes in there because its locked. So its puzzling that the mother board just died on its  own. I've seen it happen before but those odds are very low in my experience. 

[Dela7460]

I got this resolved.

The original power supply was still good; but since  I already got a new one I went ahead and swapped it.

The old power supply had a odd auto reset failure switch rather than a fuse and I just didn't want to mess with it.

 

The real problem came out to be board tension. At the bottom of the board I had a USB adaptor that plugged into the motherboard.

This adaptor sat at the very bottom of the board and held my Flash stick for Unraid. The metal housing of the adapter was making contact with the bottom of the case.

Which was fine; but it was causing the board to be under tension. 

 

For some reason a motherboard support peg (stand off) between the ram and the  SATA plugs was not exactly aligned and  was a bit cocked eyed and the bolt holding down the board to it was not the right size and was causing the board to be off kilter. This was causing a short some how.

 

Switching out the stand off and bolt and reseating everything seemed to have helped.

I still have the USB adapter pugged into the same place but I plan on changing it to one that doesn't touch the case when plugged in their.

 

Over all the sever got a good cleaning and I tossed in a 970TI that was laying around.

 

I do have 2 more sticks of ram I want to toss in it but it seems that every time I do the server freezes a week later.

I've done extended prime 95 test for over a week on each stick and can't find any faults.

 

Without fail each time those addition sticks are in the slots the server freezes with in a week.

Other wise it go's without issue for years. 

 

Even if I swap out the current sticks with the new ones it still eventually freezes.

 

Yes same brand, same everything on the ram. Way to late to RMA them. 

Board ram max is 128g  I currently have 32g in it and wanted to upgrade that to 64g.

Ram is mainly used for my VM's.

Back in the day when I ran a Ark server that thing would gobble it up.

Now I hardly run any VM's so no real reason to add the ram other than I already paid for it and its just sitting their.