Upgrade from 8.3 to 8.4 server no longer boots

Sep 20, 2024
4
1
3
Hello Proxmox Community,

Just started testing the upgrade to 8.4, first host didn't come back after patching. The system is hanging at "loading initial ramdisk" so I assumed it was a bad kernel patch. Selecting one of the prior kernels on the GRUB menu also fails with the same issue.

I'm not sure why all available kernels are dead after patching to 8.4.0.

Recovery kernels also hang at loading initial ramdisk.

Before wiping the server and starting over, I'd like to understand why the patch failed.

Hardware: HPE ProLiant DL360 Gen10, HPE SmartArray RAID, 8x 1.9TB SAS SSDs, EXT file system

Thanks
 
Before wiping the server and starting over, I'd like to understand why the patch failed.
The problem is probably not the update/patch, because an update does not touch older kernels.
Can you run the memory test? RAID status is OK?
 
At the grub menu before boot, 'e' to edit and temporarily remove the 'quiet', then hit ^X to boot - will give you more info.

A few months ago my Beelink EQR6 looked like it was hanging but GUI and ssh was still reachable by IP address, and VMs were running fine. Turned out to be AMD GPU stuff in the grub line that needed fixing.
 
So I completed a RAM test (1.5TB so it took a long while - all good) and the RAID is good. After leaving the server unattended stuck on the boot process, it completed booting.

Prior to this successful although long boot run, I had gotten into single user mode and saw that the system was hanging up on a "lockdown=none" kernel parameter flag. I took this off and rebuilt grub. The reboot after that also hung the same way so I assumed it was still broken. Now the server is online and healthy. It took approximately 20 minutes for the boot process to get past the initial image load. There are no logs indicating why the process took so long. Once the boot process got past loading in the kernel image it went pretty quick.

I'm going to patch a 2nd server of the same type and monitor it.