New Unraid server keeps freezing, crashing and disappearing from network. [6.11.5]

[NullZeroNobody]

Hi all, I am currently building my 2nd Unraid server, in order to save on energy costs. Currently, I am running:

 

Dell Precision T7500 Dual L5640, 48GB ram, GTX 1050Ti (Plex), RTX 2060 (VM)

 

And looking to move to:

 

Qnap TS-453Be, 16GB ram. (7th gen Intel Celeron J3455)

 

Previously, the Qnap has been rock solid, easily running 180 days uptime on stock software before stopping to do a system update. However, as time went on, I found the system to get more and more sluggish, and basically ran into issues with anything other than just simple file serving and pihole. Also, when I decided to shrink my Plex volume in order to add more space to another share, it decided to corrupt the entirety of my plex volume, so there’s that.

 

I purchased 4 x Seagate Exos X16 ST12000NM001G 12TB refurbished drives, to upgrade and increase storage from the previous 4 x 8 WD reds I shucked from Best buy. 3 of the 4 passed preclear, but one didn’t. After RMAing and replacing the drive, all 4 drives are good. Also purchased a PCi-e x4 m.2 adapter for a 1TB NVME cache drive. 

 

Issues I’m running into: my server will randomly reboot, under load or not. Previously I thought it was an issue with Docker network type, so I switched that to ipvlan. That seemed to do the trick for a few days, but then the reboots started happening again. 

 

I noticed that when the server dies, there is no longer any network presence from the NAS, but if I turn on the monitor it’s attached to, it shows the usual command prompt for root login, but doesn’t register keystrokes.

 

Server power usage also goes from 50-55w to 25-30w when this happens. 

Fix Common Problems finds an issue with machine check log, but that is just an issue with unidentified cpu (Intel Celeron J3455). I also cannot install mcelog since Nerdpack is depreciated, I believe. 

 

Feb 18 11:07:32 BabyBertha  mcelog: Unknown Intel CPU type family 6 model 4294967295

Feb 18 11:07:32 BabyBertha  mcelog: Kernel does not support page offline interface

Feb 18 11:07:32 BabyBertha  mcelog: Unknown Intel CPU type family 6 model 4294967295

Feb 18 11:07:32 BabyBertha  mcelog: Running trigger `unknown-error-trigger' (reporter: unknown)

 

Also getting CMCI Storm MCE  errors

 

Feb 18 12:48:32 BabyBertha kernel: mce: CMCI storm detected: switching to poll mode 

Feb 18 12:53:32 BabyBertha kernel: mce: CMCI storm subsided: switching to interrupt mode

 

Side note, I currently have both servers on the same subnet, 10.0.1.x - is this an issue? Each has it’s own assigned IP address, but I’ve read here and there that Unraid servers should be on separate subnets. Neither server is doing any DHCP or network duties, aside from running pihole. 

 

 

Any help would be appreciated.

 

Edit: attached syslog from remote server. 

babybertha-diagnostics-20230218-1342.zip

syslog-10.0.1.23.log

Edited February 18, 2023 by NullZeroNobody
Clarify Topic, Attached Syslog