Recently I have had patches hang, forcing a hard reboot. After that, there seem to be “gremlins” in the system. I have selected patches which I thought to be potentially problematic and forced reinstallation.
I am using my computer to do difficult and time-consuming work (math and physics). I would rather not spend a lot of time chasing gremlins.
One idea I have is to be able to go back to a certain date, or patch release and reapply all patches from then until the current state.
Is that currently possible without getting into the weeds of system configuration?
Is it even a good idea? I don’t know what post-installation scripts, etc. might do, or fail to do if they hang. I seems like forcing a reinstallation has fixed things in the past.
When the system stops responding to all HID and network activity, I assume it to be hung. That’s when I force a reboot. There could/should be a means of reapplying recent udates in case something has gone wrong.
About a week ago, I had updates fail, most likely due to previously failed updates. After that the system became unusable until I ran the recovery boot option multiple times and force reinstalled the kernel and systemd.
Today I had the kind of lockup that was happening frequently at the time of the failed updates. I suspect a problem has been reintroduced. I just did a zypper dup which did far more than I expected. Perhaps that will fix things.
Be aware that also for my good running system it lists some missing files and > 100 modified files.
If you have BTRFS you should have a snapshot that you can go back to that was still stable and reapply the updates there.
If you do not have that, I would just reinstall the system. I have a separate /home which makes that less painful. After that I spend ~2 hours doing some special config based on a document I have that documents (almost) all config I did in the past. In the following weeks I do the remaining config of less used functionality.
If you force a reboot during install of updates there’s a fair chance your rpmdb is corrupted. Do sudo rpm --rebuilddb reboot and see if the problems persist after a sudo zypper -vvv dup. Post output here when in doubt before taking actions.