https://bugs.kde.org/show_bug.cgi?id=525919
--- Comment #13 from Norbert Harrer <[email protected]> --- (In reply to yhl3051 from comment #12) > To begin with, I did choose the dropdown, chosen before pressing Update. > Same procedure as the failing runs. > > Nothing relevant changed here however. Current versions, same as when it was > failing: > > plasma-discover-6.7.5-1.fc44 > plasma-discover-flatpak-6.7.5-1.fc44 > plasma-workspace-6.7.5-1.fc44 > kwin-6.7.5-1.fc44 > PackageKit-1.4.0-1.fc44 > fwupd-2.1.8-1.fc44 > > The journal shows 20 consecutive Discover-triggered failures from Sep 18 to > Oct 14 08:47 EDT, then the one success at about 10:53 EDT. So the success > has a cause I could not identify. It may be timing, or some state difference > I did not capture, rather than a fix. > > Differences I could find for the successful run: > - Only 3 RPM updates, all third-party (google-chrome-stable, > google-chrome-canary, chatgpt), no Flatpak or firmware entries. > - That Discover instance was launched from a terminal with > QT_LOGGING_RULES="*discover*=true" (I was debugging another bug), which may > change timing slightly. > > I checked whether Flatpak updates in the same run correlate with the > failure: only 3 of the 20 failing runs had any, so they are not the trigger. > > At my next update I will capture the session bus with the same busctl > command you used and report the logoutAndReboot count. > > The log analysis was done with AI assistance; the data is from my own system. Thanks for checking. I tried to reproduce the conditions of your successful run: a single RPM update (plasma-workspace-wallpapers, downgraded to 6.6.4 beforehand and removed from the PackageKit cache so it had to be downloaded again), no Flatpak updates pending, and Discover started from a terminal with QT_LOGGING_RULES="*discover*=true". It still failed here, as did two earlier runs today (one without the logging rule, one with a Flatpak update in the same run). The session-bus capture of that run shows the same pattern as in comment #4: two logoutAndReboot calls from Discover with consecutive serials, this time 754 ms apart, followed by the rejected second closeWaylandWindows and resetLogout. Like you said, it might be down to timing. I'm curious what your capture shows at the next update, and whether it keeps working for you. -- You are receiving this mail because: You are watching all bug changes.
