Solution to PVE upgrade failure due to insufficient boot space
During the default installation, the space allocated by proxmox to /boot is less than 300M. The iteration of proxmox is very fast, and /boot can easily be filled up. If you are not careful, the upgrade will fail. The only way is to manually delete the old kernel.
Delete old kernel
proxmox is based on debian. Every time you use apt update && apt dist-upgrade, debian’s linux-image and pve-kernel will be automatically installed. These two items will occupy the space of /boot, so to make space, delete these two old kernel.
- Copy and use kernel
- Find old kernel
- Delete it
NOTE: If the upgrade of
proxmoxfails due to insufficient space in/boot,update-grubwill also fail, but don’t worry, every time you use the oldapt purgekernel,update-grubwill be reinstalled. Naturally, the latestkernelwill be the preferred startup option. If you are not sure, you can usegrub-set-default $menuentry_idandgrub-reboot $menuentry_idto manually set the defaultkernelfor the next startup. The above steps are still a bit troublesome, and I have to think about/bootfrom time to time. Fortunately, there are two lazy methods:
- When installing
proxmox, manually set the /boot partition to 2G, which will at least keep it clean for a while.
Reference article
uname -mrs
dpkg --list | grep "linux-image"
dpkg --list | grep "pve-kernel"
apt purge linux-image-$old_version