Dell R7515 - CPU 1 error - Intermittent Crashes

[mrtechsupport]

We've been experiencing intermittent crashes with our Proxmox server, with some occurrences happening after several days, weeks, or even months.

To address potential hardware issues, we've updated the iDRAC drivers and firmware. Additionally, we've optimized the system profile in the BIOS for performance and disabled C-State. In Proxmox, we've disabled EDAC as well.

Despite these measures, we haven't found any error logs explaining the crashes, except for the message from iDRAC indicating a CPU 1 machine check error.

Has anyone experienced the same issues? Or can suggest what things to check?


++++++++++++++++++++++++++++++++++++++++++
PVE Manager Version: pve-manager/7.1-7/df5740ad
Poweredge R7515

 

[Petros]

I had a similar issue with my R7515, but the issue was from the CPU, I think. I had an Epyc 7763, and after replacing the CPU with a new one, the issue was gone.

Don't you mind if I ask you to share a screenshot of the sensor ( CPU temperature )?

 

[mrtechsupport]

I had a similar issue with my R7515, but the issue was from the CPU, I think. I had an Epyc 7763, and after replacing the CPU with a new one, the issue was gone.

Don't you mind if I ask you to share a screenshot of the sensor ( CPU temperature )?

Hi Petros, thanks for the reply.

How were you able to confirm that it's an actual CPU issue? Was it overheating?
Btw, I have attached the screenshot.

 

[Petros]

What is currently your cpu usage? this is my temp with 30% cpu usage but I still think its way to much.
Yes, i think it was easy overheating and when cpu usage reach up to 70% easy to crash.

Please open idrac/bios => cpu and put TDP on minimum or normal if its on max and see if its still crashing but I'm afraid even that won't resolve your issue. Which CPU do you have? and what is your BIOS version?

 

[adidibau]

Hello
Did you find any solution for this I have ALOT of these servers and have exact the same problem, but im not running proxmox on then im running different OS.

 

[Petros]

Me, as well. I've bought a bunch of these servers, hah. Can you please get in touch with me directly?

I can take a look at your idrac through anydesk If you want.