HL15 Full Build - Machine Check errors

Hardware
1

I’m hoping I’ve tagged this post with the appropriate tags. I’ve got a new HL15 full build and running into some potential issues with “machine check events” getting logged. Right now these are being corrected automatically but I’m a bit leery about it.

Info about the system:
CPU: Xeon Bronze 3204
Mobo: X11SPH-nCTF
RAM: 8x 16GB DDR ECC, HMA42GR7MFR4N-TF
OS: Unraid 6.12.8

  • I already ran memtest86, passed w/o an issue
  • No errors at startup
  • Seems to start happening once I start a copy from my old NAS to this server
  • SMB mounted folder, using rsync -avzhp /location1 /location2
  • Array is formatted with ZFS
  • I’ve got an output from /var/log/mcelog but having trouble determining which module(s) it could be though I’m not even sure if mcelog is compatible with this motherboard
  • As for potential culprits I’m seeing these being logged there (can post the entire mcelog but I think I need approval first)
    CPU 0 BANK 7 TSC 9f96bc1f192
    MISC 200000c000601086 ADDR 203fdee2c0
    MCA: MEMORY CONTROLLER RD_CHANNEL0_ERR
2

The version of memtest86 included with Unraid is actually on the older side. I got the latest and greatest and 25min into the test getting some ECC errors. Hoping to figure out which slots are at fault

3

End issue was one of the sticks was indeed producing a ton of ECC correct errors. Got it swapped with another stick and its been fine for a few hours now

2 Likes