The xfsprogs side, LP: #2171360, is marked a duplicate of this bug, and that is right: the success the weekly timer reports is xfsprogs' design, so the kernel option is the only fix.
The units set SERVICE_MODE=1, and in service mode xfs_scrub returns 0 when its scrub setup fails, because "we haven't determined that there's anything wrong with the filesystem" (scrub/xfs_scrub.c, end of main(), xfsprogs tag v7.2.0); every other error returns 1. Run by hand, the same check exits 4 (operational error); with SERVICE_MODE=1 it exits 0. Measured on a throwaway XFS image with xfsprogs 6.18.0-3 and with Debian's 7.2.0-1, on 7.0.0-31-generic. So on every 26.04 kernel listed above, the timer the package enables runs every week, checks nothing, and its units end successfully: a check that cannot start never fails its unit, so no OnFailure= handler fires for it. Enabling CONFIG_XFS_ONLINE_SCRUB, the upstream default since Linux 6.18, makes the shipped timer do what it says. -- You received this bug notification because you are a member of Ubuntu Bugs, which is subscribed to Ubuntu. https://bugs.launchpad.net/bugs/2171359 Title: resolute: enable CONFIG_XFS_ONLINE_SCRUB (upstream default y since Linux 6.18) To manage notifications about this bug go to: https://bugs.launchpad.net/ubuntu/+source/linux/+bug/2171359/+subscriptions -- ubuntu-bugs mailing list [email protected] https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs
