System Fatal error on CPU - system freeze

[prototact]

@dosu Recently my system completely froze with the following being retrieved by the following command
journalctl -b -2 --priority 1
the entries were

Jul 05 12:18:41 bluefin kernel: [Hardware Error]: System Fatal error.
Jul 05 12:18:41 bluefin kernel: [Hardware Error]: CPU:1 (19:21:2) MC14_STATUS[Over|UE|MiscV|AddrV|PCC|SyndV|UECC|Deferred|Poison|Scrub]: 0xffffffffa6216620
Jul 05 12:18:41 bluefin kernel: [Hardware Error]: Error Addr: 0x0000000000000000
Jul 05 12:18:41 bluefin kernel: [Hardware Error]: IPID: 0x0000000000000000, Syndrome: 0x0000000000000000
Jul 05 12:18:41 bluefin kernel: [Hardware Error]: Bank 14 is reserved.
Jul 05 12:18:41 bluefin kernel: [Hardware Error]: cache level: RESV, tx: INSN

What can I do to avoid this in the future? I updated the system using ujust update but I am not sure it is enough.

[dosubot[bot]]

This is a hardware error, not a software/Bluefin issue. The log tells us:

• CPU family 19h, model 21h = AMD Zen 3 (Ryzen 5000 series or similar)
• MC14_STATUS with UE (Uncorrectable Error), UECC, Poison, Scrub = an uncorrectable ECC memory error was detected by the CPU's memory controller
• PCC (Processor Context Corrupt) = the CPU determined its own state was no longer trustworthy, so it halted — hence the freeze

ujust update keeps your kernel and firmware current, which is good practice, but it won't fix failing hardware. Here's what you can do:

1. Update your BIOS/UEFI firmware — AMD has shipped microcode and AGESA updates that improve memory error handling. Check your motherboard vendor's support page for the latest BIOS.

2. Run a memory test — Boot into memtest86+ (select it from GRUB or your UEFI boot menu) and let it run several passes. An uncorrectable ECC error strongly suggests a faulty DIMM or memory controller issue.

3. Check physical RAM — If you have multiple DIMMs, try removing one at a time to isolate the bad stick. Also reseat them to rule out a contact issue.

4. Monitor for recurrence — After your next boot, run:

   journalctl -b --priority=3 | grep -i "hardware error"
   

   A one-off event can be a cosmic ray bit-flip (rare but real). Repeated errors point to a hardware defect.

5. File a bug report for tracking — If it recurs, you can run ujust report to collect diagnostics (it auto-scrubs PII) and file it with the Bluefin team for visibility [1].

Bottom line: This is almost certainly a RAM or CPU hardware fault. No amount of software updates will prevent it if the hardware is degrading. If memtest shows errors, replace the affected DIMM.

Docs: https://docs.projectbluefin.io

To reply, just mention @dosu.

---

Share context across your team and agents. Try Dosu.