Hello, we are facing a very peculiar issue with Proxmox.
We have these servers Intel SCL1530CLR from ~2006 (with two Intel Xeon E5420 cpus, 8GB DDR2 and S5000VCL motherboard ). We have been trying to install Proxmox VE into them for a project, but every time we open the installer there is a big chance of it locking up and not progressing, but sometimes the installer do go through. We have attempted to diagnose the installer also locking up, to no avail.
These were two different attempts at booting the Proxmox VE installer, it gets stuck at different points.
This also happens with other Linux distribution installers such as Debian and Ubuntu. We have gone around that by attempting to install several times until the installer finally loads. After install, every time we boot the machines up, they get stuck at a random point at boot. If left for long enough (and removing quiet from the options at GRUB) they will display this thing:
Looking over on Google, this post from AskUbuntu tells us to add as parameters noapic and nomodeset. Doing this will not have any effect, as the servers keep doing what stated above. Testing other operating systems, stock Debian 12 works. Another measure we have tried is to install the Debian kernel 6.1.0-21 onto Proxmox and then boot it from GRUB. It is not the best idea, but doing this will make the system boot all the time. Another thing we have tried is to install the intel-microcode from this suggestion, to no difference. Another thing we have tried is to install the Debian Sid kernel into Proxmox, which is also not beautiful, since the package manager forced us to uninstall proxmox-ve, but we wanted to know if it was down to the newer kernel versions, to which it also was able to boot.
Finally what we have tried today was to debug the boot process using the serial console during the system boot with the kernel parameter console=ttyS0,112500n8, to which the system does boot every time we have tried with no issue using the stock pve kernel, and we were not able to replicate the bug with this option set.
Any ideas as to what would be happening that would prevent the system from booting unless these specific scenarios? If it also helps, we can attach pictures of when Debian locks up during the install phase, which does give some more information.
Also attached are the logs of the kernel once the system boots correctly (from dmesg).
Thanks in advance.
We have these servers Intel SCL1530CLR from ~2006 (with two Intel Xeon E5420 cpus, 8GB DDR2 and S5000VCL motherboard ). We have been trying to install Proxmox VE into them for a project, but every time we open the installer there is a big chance of it locking up and not progressing, but sometimes the installer do go through. We have attempted to diagnose the installer also locking up, to no avail.
These were two different attempts at booting the Proxmox VE installer, it gets stuck at different points.
This also happens with other Linux distribution installers such as Debian and Ubuntu. We have gone around that by attempting to install several times until the installer finally loads. After install, every time we boot the machines up, they get stuck at a random point at boot. If left for long enough (and removing quiet from the options at GRUB) they will display this thing:
Looking over on Google, this post from AskUbuntu tells us to add as parameters noapic and nomodeset. Doing this will not have any effect, as the servers keep doing what stated above. Testing other operating systems, stock Debian 12 works. Another measure we have tried is to install the Debian kernel 6.1.0-21 onto Proxmox and then boot it from GRUB. It is not the best idea, but doing this will make the system boot all the time. Another thing we have tried is to install the intel-microcode from this suggestion, to no difference. Another thing we have tried is to install the Debian Sid kernel into Proxmox, which is also not beautiful, since the package manager forced us to uninstall proxmox-ve, but we wanted to know if it was down to the newer kernel versions, to which it also was able to boot.
Finally what we have tried today was to debug the boot process using the serial console during the system boot with the kernel parameter console=ttyS0,112500n8, to which the system does boot every time we have tried with no issue using the stock pve kernel, and we were not able to replicate the bug with this option set.
Any ideas as to what would be happening that would prevent the system from booting unless these specific scenarios? If it also helps, we can attach pictures of when Debian locks up during the install phase, which does give some more information.
Also attached are the logs of the kernel once the system boots correctly (from dmesg).
Thanks in advance.