/forcefsck is obsolete

Modern file systems are expected to determine themselves if fsck is
needed. XFS, ext4, btrfs, all can do log replay at mount time and that
fixes the vast majority of problems. On XFS and Btrfs, if log replay
has problems, mount will fail and it's expected the user manually runs

So right off the bat, /forcefsck would apply only to ext4 on Fedora.
But there's another gotcha which is the file system has to be
consistent enough to mount ro in order to find the forcefsck hint. So
it's kinda unreliable. Next it required the use of preen on a read
only mounted volume, which upstream doesn't recommend. From e2fsck man

"Note that in general it is not safe to run e2fsck on mounted
filesystems.  The only exception is if the -n option is specified"

In a somewhat recent systemd, it just pre-emptively runs the generic
fsck on rootfs before ro mount. On ext4 this causes e2fsck to get run
prior to ro mount which if the file system state is clean, it's a no
op. On XFS and Btrfs fsck is always a no op.

Chris Murphy
users mailing list -- users@lists.fedoraproject.org
To unsubscribe send an email to users-le...@lists.fedoraproject.org

Reply via email to