Closed jlebon closed 6 months ago
I think this touches on a paper cut I thought would be handled by https://github.com/coreos/zincati/issues/498 when we finally implemented that.
As is it improves the state, but I think users should have a documented way to do this too.
I think this touches on a paper cut I thought would be handled by #498 when we finally implemented that.
for example, #498 could handle the "I'm in an interactive session, an update is available, and I want to ignore update strategy (i.e. if your node isn't configured to update until the weekend)" case.
I think this touches on a paper cut I thought would be handled by #498 when we finally implemented that.
Yeah, thanks for calling out #498. I agree it'd mostly solve that case, though OTOH it's also a lot more work. I found this useful for debugging https://github.com/coreos/fedora-coreos-tracker/issues/1691 and it took a few minutes to type it out. We can nuke it when #498 happens.
An issue as old as ~time~Fedora CoreOS is that when you're testing something related to updates, you want Zincati to reboot but because
cosa run
gives you an interactive session and you can't not have a session without losing the VM, you're stuck waiting for the 10 minutes grace period to expire.Just add a hidden knob for this to make that scenario less painful.
The API is pretty much: when you're ready to have Zincati reboot, just
touch /run/zincati/override-interactive-check
.