System broken by update

[alex769]

Since recently my installation detaches itself from the network after 1-2 days and I have to hard reset the machine. The problem exists since an automated update on October 29th.
Does anyone know hot to fix this or how to revert back to a proper kernel version?

 

[Maximiliano]

Hello, the command:

last reboot -F -n 20

will tell you the last 20 booted kernels. If you want to boot from a specified kernel version, please take a look at our documentation [1].

Having said that, is there anything in the system logs pointing why you lose network connectivity?

[1] https://pve.proxmox.com/pve-docs/pve-admin-guide.html#sysboot_kernel_pin

 

[alex769]

I looked through the logs and each time they just stop at 3 o'clock and some minutes. Looks like the system crashes. One during backup and once during emailing to root@pam
Then there are no more entries until the reboot remark.

 

[gfngfn256]

One during backup and once during emailing to root@pam

IDK your HW, but assuming the email is a backup notification, we can assume that the crash is happening at either the time of the backup or just after it. Since you show no other log entries after this - we can assume you have a hard crash where all components of the system are crashed. So that backup is causing a hard crash. I would check in the following order; Cooling, RAM & PSU.

 

[alex769]

One of the VMs runs TrueNAS, which produces a good log. There I could not see any correlation to load or exact time. Once the machine died shortly after 2h with hardly any load on the system and the other time it died shortly after 3h during backup, when the trueNAS VM was under significant load. If it was a cooling issue, it would rather die during the day, when I work on it and there is more going on. Power supply is not likely.
Is there a way to run test cycle on the RAM?

 

[Kingneutron]

https://www.memtest.org/

 

[SamTzu]

Try the new kernel Linux 6.8.12-4-pve (2024-11-06T15:04Z)

 

[gfngfn256]

You mention in your original post:

The problem exists since an automated update on October 29th.

1. How long was the system running/working correctly before these updates?

2. What automated update to you refer to? PVE (by default) doesn't automatically update itself.

3. What convinces you that it is linked to these updates?