Zoom is vital to my job this month and prior to an update last week I had the openSUSE version of Zoom's RPM installed and working fine.
I updated my Tumbleweed installation to openSUSE-20240704-0 last week, after which Zoom started crashing when sharing a screen. There was a message in the logs about the library libqt5qml.so and I thought I could fix this by backing out either the update for the libQtQuick5 package in particular, or just booting from the pre-update snapshot.
To make a long story short, I ultimately installed the Zoom Flatpak and resolved to get back to this when I had a bit more time.
My question - Can people suggest the right way in openSUSE Tumbleweed to handle the situation where an update breaks something on the system?
Assuming libQtQuick5 was the updated package that was at fault here, is there a way I could have downgraded just that package? Would booting from the pre-update snapshot and then just carrying on with my week have been a reasonable way to proceed?
To be clear - I'm not so much concerned about Zoom, I'm more curious about how to use the openSUSE Tumbleweed tools to recover from updates that cause problems.
If a tumbleweed update causes issues, then rollback to a previous snapshot where the issue was not present. You should: reboot into the preupdate snapshot, which will be in read only mode; then launch zoom and check it works without error; then in a terminal enter sudo snapper rollback, which creates a read write copy of the good snapshot; reboot and your computer should be as if you never did the update.
The next question is about updating again in the future and if zoom will break. If you update again in say a month, the issue may not occur. It's hard to say. Otherwise, if you can figure out which library packages caused the problem, you can lock them in yast so that they do not update on zypper dup.
I had a similar issue with zoom in the last few years. If zoom stability is essential for you, just use the flatpak. The web client is an option too.
Had an issue many many years ago with TW where X11 was completely broken on my system. I just did a btrfs roll back to prior to the update and then updated like a month later
I think the correct way would be to boot from a previous snapshot and roll back to that. No idea what could go wrong when you're staying in a snapshot. I'd mostly be afraid that I'd forget that I'm on a snapshot and find myself with the error again on next boot.
Snapshots are read only. Best plan is to rollback to a snapshot you think works, test it and if all is good use sudo snapper rollback to make the current snapshot the default. I usually reboot at that point too, not sure if it's necessary though.
Most people have answered doing a rollback is the best way. I usually find some updates break things then later updates dont have the issues.
But I wanted to add if you go in to yast snapshots and double click a snapshot you can actually select specific changes to rollback via checkboxes. I've not tried this yet though because of dependencies and whatnot
I think you did alright by resorting to flatpak. When it comes to closed source applications, it's not uncommon for problems like this to appear every so often.
Using the flatpak is the right answer. Seriously though flatpaks are separate from the base system so this isn't an issue.
The bigger issue is that you are running tumbleweed on a production critical machine. If you want to run it on a personal machine that isn't critical that's fine but for production stick with well tested. Things will break and its best to stick with slow and stable. Think Linux Mint or Debian with Flatpak apps.