Recently our unraid server was hit by a power surge. Despite everything we have being plugged into surge protectors it blew out a cable modem, switch, laptop and our unraid server. Prior to this, the server had performed flawlessly for a year.

 

I replaced the MB and CPU and rebuilt the server. Everything was fine for a week, but over the last two days the unraid server has crashed twice.

 

After each crash, when I restart the server the parity is checked and no errors have been found.

 

Where do I begin trying to debug this? For example, where are the logs on the server I can examine to try and diagnose the crash?

 

I replaced the MB and CPU and rebuilt the server.

 

Did you also replace the memory and the power supply?

 

 

After each crash, when I restart the server the parity is checked and no errors have been found.

 

This is normal -- anytime UnRAID does an "unclean" shutdown, it will automatically check parity on the next restart.    Your server is shutting down abruptly ... so that triggers this action.

 

 

Where do I begin trying to debug this?

 

Boot the system and select MemTest instead of UnRAID.      Let it run for at least 6 full passes (overnight).    If it encounters ANY errors, you need to replace the failed module -- if you have more than one, you'll have to remove one; repeat the test;  then insert the next one; repeat the test; etc. until you isolate which module has failed.

 

If there are no memory issues, then I'd replace the power supply [especially if you didn't do that as part of the rebuild].

 

 

I did NOT replace power supply or memory

 

Am running memtest right now and will let the forum know what happens

 

Thanks for the input

There's an excellent chance the issue is one of those two things then (as I had assumed was the case, but wanted to confirm you hadn't replaced them).

 

One other thing:  A UPS is a very good idea -- it not only provides superb surge protection;  but also ensures your system never encounters an "unclean" shutdown => if power is lost for more than a specified time (typically you'd want to set this for 5-10 minutes), the system will be gracefully shut down.    I consider a UPS a mandatory accessory for any computer -- but especially for a server.