Very strange, my Supermicro server that had been up for just over 7 days just up and rebooted. I was sitting next to it and from my peripheral vision I saw something flash on the monitor that got my attention but it went by to fast, next thing I knew the server was posting. Really weird, there had been nothing to indicate anything was wrong in the logs or anything I had been checking now and then just to see. I'll attach my logs although there obviously isn't much in them. Its doing a parity check because it detected an unclean shutdown.

 

 

tower-diagnostics-20170207-1833.zip

Do you have an UPS? Is it in good condition?

Yes, SmartUPS 1500 and yes its in good condition.

I was having random reboots for awhile due to Plex being run as a plugin and not in a docker.  The random reboots went away once I switched to using a docker for plex.  I have no idea if you are running plex but just chiming in as to what caused this for me.

A failing PSU can do this as well, actually a number of hardware components (obviously), though, not sure if you did the PSU conversion on the SuperMicro or if you're still running the redundant stock PSU's.

Thanks, I have been running plex as a docker for sometime now, well before I migrated to the Supermicro chassis. Also, I am running the stock dual redundant power supplies that came with the Supermicro chassis.

*Nevermind, ninja add or I simply didn't see the attachment before.

They were there the whole time 

I figured out what the cause of my random reboots, I have a bad stick of RAM. After every one of these events an OEM System Boot Event is recorded. Wonder if I can run the system without this DIMM or if I have to replace it?

 

841 2017-02-16 07:05:29 Mmry ECC Sensor Memory Uncorrectable memory error ; Logical Rank 0 ; DIMM = C2

 

832 2017-02-07 23:16:10 Mmry ECC Sensor Memory Uncorrectable memory error ; Logical Rank 0 ; DIMM = C2

 

821 2017-01-23 04:48:54 Mmry ECC Sensor Memory Uncorrectable memory error ; Logical Rank 0 ; DIMM = C2

 

So it did turn out to be a memory problem after all! Was it MemTest version 7 that revealed it?

 

No actually, I remembered when I had the IPMI plugin installed I was able to access a log of events, it was in this log I found these memory events. I just removed the faulty stick of ram and the server is booting now. Time will tell if that was it, but I am pretty sure it must have been.