On a CentOS 6 host I run Rocky 9 VM’s without any problems. Now I tried several nethods to initialize a RL 10 VM, without success. I tried virt-install with -l (location …/10/BaseOS/x86_64/os/ ), with -c (boot.iso), but all attempts end up in “no bootable disk”. I also tried a generic cloud image (converted from qcow2 v3 to qcow2 v2), resulting in the same error. Is it really impossible to run a RL 10 VM on CentOS 6?
I know the status of CentOS 6 (it’s an old host I still use for devel VM’s) and I do not consider my issue as a bug of RL 10 or so. Still, I’m curious to know what’s causing this and if there maybe is a workaround.
Could well be due to changes and advancement in KVM VM’s that CentOS 6 simply doesn’t have. It would make no sense for Rocky to provide compatible KVM images for CentOS 6 when it’s already a deprecated platform. I expect they are built for current technologies.
Maybe the VM is created for MBR when UEFI is required, maybe CentOS6 doesn’t have the UEFI requirements to boot it. Could be pretty much anything.
I would expect however for it to work the other way around, eg: install a CentOS 6 VM on a Rocky 10 KVM host. Old will mostly work on new systems, but new will pretty much be problematic to run on old systems so is to be expected. I doubt very much there will be a workaround, and TBH nobody would bother spending the time on it simply because it’s old deprecated and EOL.
Probably better, migrate/backup your VM’s, and then clean install this CentOS 6 machine with something newer.