Skip site navigation (1)Skip section navigation (2)
Date:      Fri, 02 Jul 2021 19:36:25 +0000
From:      bugzilla-noreply@freebsd.org
To:        fs@FreeBSD.org
Subject:   [Bug 256936] Buggy filesystem detected - message wrongly triggered by FUSE filesystems
Message-ID:  <bug-256936-3630-3ZaY03HFkm@https.bugs.freebsd.org/bugzilla/>
In-Reply-To: <bug-256936-3630@https.bugs.freebsd.org/bugzilla/>
References:  <bug-256936-3630@https.bugs.freebsd.org/bugzilla/>

next in thread | previous in thread | raw e-mail | index | archive | help
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=3D256936

--- Comment #2 from Alan Somers <asomers@FreeBSD.org> ---
If the FUSE daemon tells the kernel that it may cache dirty data, but then =
the
daemon tells the kernel that the file has changed while it was dirty, then =
data
has been lost.  Data corruption is a bug.

MooseFS's hands are somewhat tied as it still uses fusefs-libs2.  At that
protocol level there really aren't sufficient operations for the server to
manage the kernel's cache.  But sshfs uses fusefs-libs3, so it ought to be =
able
to fix the problem.

If it's possible to prove that no data corruption occurred and yet the kern=
el
still prints this error message, then that would constitute a bug in the
kernel, and I would fix it.  The most obvious way for that to happen is if =
you
set vfs.fusefs.data_cache_mode=3D0, completely disabling the cache.  But ev=
en if
data_cache_mode=3D1, it wouldn't surprise me if there are cases where the w=
arning
is printed inappropriately; it's complicated.  Especially because file
attributes can be cached independently of their data.

--=20
You are receiving this mail because:
You are the assignee for the bug.=



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?bug-256936-3630-3ZaY03HFkm>