It seems reasonable that setting TZ=UTC0 for date --iso-8601 could help (as that's where a time zone is relevant, the +%s conversion it is not relevant for).
But then this causes unwanted behaviour for people far off UTC, I guess, so it's not a solution - It's entirely unclear to me whether a solution exists given that we allow arbitrary intervals in settings. Always-on machines (or well, always connected machines) should be configured with the interval set to 0 (disabling timestamp checks), and the timer should be configured to run at the appropriate times. To resolve issues with skipped runs elsewhere, we really need to resolve the issue that systemd can't reschedule failed timer services, and then remove the options inside apt.conf, as having dynamic timers in systemd and then a fixed check inside the script does not really work (hence why we annoyingly run it twice a day to give it a chance to fixup time skew). ** Changed in: apt (Ubuntu) Status: Confirmed => Won't Fix ** Changed in: apt (Ubuntu) Status: Won't Fix => Triaged -- You received this bug notification because you are a member of Ubuntu Bugs, which is subscribed to Ubuntu. https://bugs.launchpad.net/bugs/1824088 Title: unattended upgrade ran one day after schedule To manage notifications about this bug go to: https://bugs.launchpad.net/ubuntu/+source/apt/+bug/1824088/+subscriptions -- ubuntu-bugs mailing list ubuntu-bugs@lists.ubuntu.com https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs