[SOLVED] PVE Fails to load past 'Loading initial ramdisk...'

[RustyWalls]

Hi!
This morning my PVE server stopped booting correctly, seems like this happened after a restart or something on its own as it's on a UPS and I never restart it without ensuring it works after. I've followed some other posts on this same issue but none have resolved it.

I believe im using PVE 6.2-1 but I'm unsure if I may have updated it since first setting it up, I set up the server around a year ago.
It's using Linux 5.4.140-1-pve as reported by grub.

The boot device is a 2TB NVME drive.
Running a ryzen 7 3700x on Gigabyte b450m DS3H motherboard running BIOS version F50 32GB of ram.

I believe this may be caused by IOMMU as I was messing with that a few weeks back but never had issues then.
I've tried editing the grub and removing the 'quiet' option but still hangs at 'Loading initial ramdisk...'.
I also changed 'amd_iommu=true' to 'amd_iommu=false' then did Ctrl+X but still hangs.
I've also tried disabling SVM in bios as suggested in one post but still hangs at 'Loading initial ramdisk...'.
I've also changed IOMMU in bios from 'Auto' to 'enabled' and 'disabled' but no change, still hangs at 'Loading initial ramdisk...'.

I also tried to boot into rescue boot using a USB with PVE 7.1-2 but that froze at 'VFIO'

I have some experience with Linux but not enough to fix this issue. Thank you to anyone who can help!

 

[Neobin]

I also changed 'amd_iommu=true' to 'amd_iommu=false' then did Ctrl+X but still hangs.

I can only say, that those parameters are not existing.
amd_iommu is on per default.
So the only valid option in this specific regard is: amd_iommu=off [1].

I also tried to boot into rescue boot using a USB with PVE 7.1-2 but that froze at 'VFIO'

Did you try with the PVE 6.4-ISO [2] too?

But otherwise no idea on your problem, sorry.

[1] https://www.kernel.org/doc/html/v5.4/admin-guide/kernel-parameters.html (For your 5.4 kernel.)
[2] https://proxmox.com/en/downloads/item/proxmox-ve-6-4-iso-installer

 

[RustyWalls]

I can only say, that those parameters are not existing.
amd_iommu is on per default.
So the only valid option in this specific regard is: amd_iommu=off [1].



Did you try with the PVE 6.4-ISO [2] too?

But otherwise no idea on your problem, sorry.

[1] https://www.kernel.org/doc/html/v5.4/admin-guide/kernel-parameters.html (For your 5.4 kernel.)
[2] https://proxmox.com/en/downloads/item/proxmox-ve-6-4-iso-installer

What a weird situation.
I went ahead and tried using a PVE 6.4 iso too but that got stuck in boot as well. Afterward, I made a backup of my boot drive using Clonezilla, then uninstalled the PVE 6.4 drive.
Went ahead and installed PVE 7.2 onto a new drive. Booted perfectly fine into PVE 7.2. reinstalled the PVE 6.4 drive to access its files, and I also ensured the boot order was correct-- 7.2 first then 6.4 drive after.

However, it seems to somehow booted into 6.4 perfectly fine, I logged onto the web terminal and it threw 'virtualization not enabled in bios' error, I went ahead and turned off the system and confirmed it was enabled (yup still enabled even before I did this whole process), and removed the 7.2 PVE drive. Reboot and boots perfectly fine back into 6.4.

I'm still at a loss as to how this happened or what even occurred, I was able to successfully clone the drive before I ever touched it. But it seems my issue was solved by simply reseating the boot drive- unsure if this works for anyone else who encounters an issue similar to mine.

 

[Uplore]

I found this post while trying to troubleshoot my Proxmox 7.3 master node that would consistently boot loop after reaching the 'Loading initial ramdisk...' point. The device itself is a 'Topton Fanless Mini PC Intel Core i5 10210U 6 LAN i211AT'. After a lot of effort troubleshooting RAM, BIOS battery, msata drive etc. I discovered that the external power brick that was powering the device (DC 19v 4.74A) had failed in such a way that low power operations were still possible, such as the initial POST, accessing BIOS, loading GRUB, and reaching the initial Proxmox boot menu, but any subsequent 'high power' action, such as attempting to actually load Proxmox, undertake a memory test etc. immediately resulted in a reboot. It seems like this is pretty rare, but if someone finds this post in the future and are using a NUC / mini PC with external power brick, try a different one early in your troubleshooting. This would have save me a lot of time and effort. Good luck!