Problem upgrading from 4.18.0-372.32 -> 4.18.0-425.3

Hi,

I have a few VM’s on a 2008-R2 hyper-v host which are happy when running 4.18.0.372 however after the recent patch to 425 they all refuse to boot

Timed out waiting for device dev-mapper\x2dswap.device

seems to be the first error I see, it then waits a while with dracut messages before getting to the emergency mode boot.

I think it’s got something to do with 2008-R2 as it seems ok on a different test box but I’m stuck with 2008-R2 for this particular network and won’t be upgrading it until next year.

Anyone else seen the same or got any ideas ? (I’ve just done a clean minimal install from Rocky-8.6-x86_64-minimal.iso and done a dnf upgrade then rebooted to the same state.

Thanks for any suggestions

If the issue is in the new content, then clean boot with Rocky-8.7-x86_64-minimal.iso should fail too. Does it?

Just tried booting from that CD and doing a repair and then a fresh install. Both cases give results similar to the updated kernet with no Hard Disks found to install to.

That sounds like the real issue.

If one had unsupported hardware and had installed an additional driver for 8.6 that fails to run with 8.7, then we had something to look at, but you have VM and presumably the hyper-v presents something supported?

It’s all supported with 8.6 (kerne; 4.18.0-372 is labeled 8.7 and that works too) it’s only become apparent that it’s an issue with kernel 4.18.0-425.3 .

That version does still work on newer versions of Hyper-V but the vm’s I have can’t move to a different host at the moment.

All kernels released for RHEL 8.6 (and hence Rocky 8.6) were version 4.18.0-372.*
All kernels released for RHEL 8.7 (and hence Rocky 8.7) will have version 4.18.0-425.*
There tend to be feature changes in each point release (which is why the Y in 4.18.0-Y changes).
There should be no removal of features, but changes can reveal issues.

MS does not list 2008-R2 in Supported CentOS and Red Hat Enterprise Linux virtual machines on Hyper-V | Microsoft Learn

It seems to think there are no hard drives. Most hosts have some kind of tool where you can view the devices attached to the vm guest. e.g. can you see the storage is attached in 8.6 and then in 8.7?

The other option is to boot the guest into single user emergency mode, e.g. from the grub menu, press ‘e’ and then tell it you want emergency targer, then see if there are any visible hard drives.