Since this applies to any time automatic checks are requested, perhaps a better message would be:

XFS checks always skipped when run non-interactively

Or perhaps the more informative:

XFS check skipped, use 'xfs_repair -n' if a check is required

The behavior seems well documented, this is a low priority fix to make things clearer at the system message level.

Initscripts invoke fsck.$FS any time the filesystem was not cleanly unmounted; this is not required for journaling filesystems, and so fsck.xfs is there simply as a no-op wrapper to satisfy init.   It has printed this message since 2006 ...

The manpage does also explains what it does and why.

If it's manually invoked, it does give more helpful information:

+       echo "If you wish to check the consistency of an XFS filesystem or"
+       echo "repair a damaged filesystem, see xfs_check(8) and xfs_repair(8)."

That said, we do need a more reliable way to actually invoke a boot-time fsck if desired; if/when that gets fixed it'd make sense to audit all messages for clarity.

> we do need a more reliable way to actually invoke a boot-time fsck if desired

I'm not even sure this is needed for XFS. The mount attempt is probably sufficient, given its different architecture from ext[234].

The documentation is pretty clear-- this bug report was merely to suggest expanding that clarity into the actual message displayed on screen.

Honestly, this is a discussion that should happen upstream, it's a message that has existed for years in RHEL, and it won't be changed in RHEL7 at this point.  I'm open to changing the message, but it doesn't need to happen in RHEL7.

Thanks,
-Eric

https://bugzilla.kernel.org/show_bug.cgi?id=204997 has been opened for this issue. I've also included a suggested trivial patch to improve the error message.