CPU errors AMD EPYC 7502P

[achirkov]

Hello, I have a server (on hetzner) with AMD EPYC 7502P CPU.
Matherboard:

    Product Name: KRPA-U16 Series
    Version: Rev 1.xx

     # pveversion

    pve-manager/8.2.7/3e0176e6bb2ade3b (running kernel: 6.8.12-4-pve)

I am currently running 3 lxc containers (Ubuntu 20.04) and I am starting to get errors in dmesg:

[153996.660148] [Hardware Error]: CPU:0 (17:31:0) MC17_STATUS[Over|CE|MiscV|AddrV|-|-|SyndV|CECC|-|-|-]: 0xdc2040000000011b
[153996.660683] [Hardware Error]: Error Addr: 0x00000007b9493840
[153996.661203] [Hardware Error]: IPID: 0x0000009600050f00, Syndrome: 0x0b000b000a801202
[153996.661730] [Hardware Error]: Unified Memory Controller Ext. Error Code: 0
[153996.661835] EDAC MC0: 1 CE on mc#0csrow#2channel#0 (csrow:2 channel:0 page:0x1f3524e offset:0x40 grain:64 syndrome:0xb00)
[153996.662875] [Hardware Error]: cache level: L3/GEN, tx: GEN, mem-tx: RD
[154324.297911] mce: [Hardware Error]: Machine check events logged
[154324.298839] [Hardware Error]: Corrected error, no action required.

Installed all latest versions of packages in apt, amd64-microcode is installed. Maybe I missed something? Or the server has a hardware problem?

 

[cheiss]

HI,

Or the server has a hardware problem?

Most likely, at least the messages seem to indicate some faulty RAM.

I'd suggest running memtest86+ for a bit and testing it through, that is never a bad idea. You can either run it directly from the boot menu or using the ISO - you'll need to disable Secure Boot though for it, if you have than enabled on your server.

 

[achirkov]

I tested the memory, everything is ok with it.

 

[gfngfn256]

It looks HW related. Have you checked for latest BIOS version?
If RAM is in fact good, you are probably looking at MB/CPU.
How long have you been running this server?

 

[achirkov]

The server was rented a week ago from Hetzner. I thought the problem was in the system packages, but since the problem is hardware, I'll write to their technical support. Thanks for your help.