I wanted to install Rocky 9 on a VPS, but the hosting company only provided a Rocky 8 iso image.
I tried various methods of install using the VPS rescue console, but there was not enough rescue disk space to hold any useful installation media.
I eventually hit on the idea of using the Rocky 8 iso and going through the normal installation procedure, but specified a minimal netinstall, pointing at the Rocky 9 repo Index of /pub/rocky/9/BaseOS/x86_64/os/ .
That actually seems to have worked and when the install had finished, I rebooted the VPS from its disk drive, and I have a running Rocky 9 installation.
I would like to know if there are likely to be any downstream issues arising from this installation method?
Wow, so, apart from this method likely being entirely unsupported (and understandably so), I am both impressed by your ingenuity–and the fact that it seems to be working, at least so far.
As this is the first I’ve heard of anyone doing this, I’m unsure what, if any, issues may arise post-install, having used this method. That said, if you have a normally-installed RL9 system elsewhere then I would compare at least a few things between the two as a basic sanity check, such as:
contents of /etc/redhat-release match
enabled package repos match (‘dnf repolist enabled’)
verify functionality of typical dnf operations (install, update, remove, etc.)
fully-updated RL9 systems have package versions that match (i.e. kernel, systemd, etc.)
If all that looks OK, then maybe this sort of install actually is an option in situations such as the one you were in with that VPS?
Again, congrats on coming up with an alternative approach in that environment. Hope it’s also stable for you, long-term, as well.
