https://bugzilla.suse.com/show_bug.cgi?id=1181606 https://bugzilla.suse.com/show_bug.cgi?id=1181606#c6 Fabian Vogt changed: What |Removed |Added ---------------------------------------------------------------------------- Flags|needinfo?(fvogt@suse.com) | --- Comment #6 from Fabian Vogt --- (In reply to Lukas Ocilka from comment #5)

That's a good question, Justin, but this has nothing to do with your laptop. The whole Live Installer is THE unexpected things. I wrote "unexpected and thus untested environment". YaST team does not develop Live Installer -- it's based on something we DO develop though, but it's been designed for special and well-controlled installation system where no other application can run at the same time.

Although it may seem to be rather easy to adjust that part of code to handle this situation better, it's not the only possible race condition that can happen.

But let's get back to your question. This race condition happens when someone starts the Live Installer while PackageKit still holds the lock for libzypp. The behavior can be simulated by, e.g., running `zypper patch` (there need to be applicable patches though) and zypper needs to be kept running, then Live Installer started. I don't think that openQA does that, so, untested. BTW, thanks for the bugreport including logs.

openQA actually seems to do the opposite - it explicitly disables packagekitd.service before starting the installer. So that actually hid the bug, apparently the issue was already known. I never hit it in my tests though.

Fabian, I guess you can take over from here, right?

Yes, once Justin replied to my question. If that works, the fix should be just stopping packagekit in start-install.sh. -- You are receiving this mail because: You are on the CC list for the bug.