Re: Might NOT be a Debian problem
On 9/24/25 17:13, Van Snyder wrote:
This might NOT be a Debian problem but this list almost surely has
knowledgeable correspondents, so here goes…
I upgraded my computer with a new MSI Z90 board,
I do not see a "Z90" motherboard on the MSI web site (?). Please
provide a URL.
Is your motherboard firmware version compatible with the rest of your
hardware?
Have you reset the motherboard firmware Setup settings to factory
defaults, and checked every single setting in case there is something
you must set otherwise?
Intel I9-14900K,
Intel says:
https://www.intel.com/content/www/us/en/products/sku/236773/intel-core-i9-processor-14900k-36m-cache-up-to-6-00-ghz/specifications.html
Launch Date Q4'23
Not the newest, but Debian Stable is always behind and might be missing
some microcode patch. Perhaps you should try backports, Testing, or Sid.
32 GB,
Is the memory on the motherboard manufacturer's qualified memory list?
Intel says:
ECC Memory Supported Yes
Does the motherboard support ECC memory?
Is your memory ECC? Cosmic ray bit flips are real. I prefer computers
with ECC memory.
Have you run memtest86+ for 24+ hours under varying temperatures?
1 TB NVME drive.
Any tests?
Debian 12 started crashing — display virgo intacto, but mouse frozen,
keyboard frozen with num lock and scroll lock lamps flashing,
The flashing lights could be a diagnostic code, or a bug symptom. Have
you RTFM, STFW, and/or contacted the motherboard manufacturer's
technical support?
can't SSH to it, can't HTTP to it,….
It sounds like the kernel is crashing.
Any clues when you reboot? /var/log? systemd/ journalctl?
So I upgraded to Debian 13. Same thing.
Again, Testing or Sid.
I put in the memory sticks one at a time. Crashed with either one
alone.
I assumed it's unlikely they are both defective.
Are the memory modules a matched pair from the same production run?
They should be. So, they should both work or both fail.
I replaced my antique NVidia K2200 with a new Radeon RX580. Same
problem. So that wasn't it.
I agree that the video card can be eliminated for now.
I sent the board back to MSI and they returned it to me without
comment. Did they re-cap it? Did they test it? Who knows. Problem
persists.
Tried a new power supply. Problem persists.
An inexpensive power supply tester is a worthwhile investment.
So is an anti-static wrist strap. Electrostatic discharge, however
small, is bad; very bad.
I thought Firefox might be the culprit, but it crashed a few minutes
ago with Firefox not running. Also GKrellM is suddenly getting seg
faults but I was getting crashes without it running.
Core 0 temperature varies from 27.8°C to 53.0°C but Intel says 70°C
ought to be OK. So it's probably not the heat sink or fans.
I agree that CPU cooling can be eliminated for now.
That seems to leave only the Intel I9-14900K. Is that a likely culprit?
Trying out a replacement would be an expensive experiment.
Without suitable test equipment or software diagnostics, the only option
is spare hardware and A-B testing. You have chosen fairly recent
high-end components, so finding inexpensive compatible spares is not
going to be easy.
David
Reply to: