Skip site navigation (1)Skip section navigation (2)
Date:      Thu, 10 Mar 2005 13:05:48 +0100
From:      Jeremie Le Hen <jeremie@le-hen.org>
To:        Denis Shaposhnikov <dsh@vlink.ru>
Cc:        freebsd-current@FreeBSD.ORG
Subject:   Re: unionfs 5.4
Message-ID:  <20050310120548.GL34822@obiwan.tataz.chchile.org>
In-Reply-To: <87r7it18fh.fsf@neva.vlink.ru>
References:  <87is46kzk1.fsf@neva.vlink.ru> <41C26F23F7DF023CB3DF35C5@cc-171.int.t-online.fr> <87sm3ajj8s.fsf@neva.vlink.ru> <20050305151903.GC26240@hub.freebsd.org> <87r7it18fh.fsf@neva.vlink.ru>

next in thread | previous in thread | raw e-mail | index | archive | help
> And if unionfs panic the system on 5.4 too, I think it can't be STABLE
> at all.
> 
> BTW, from man mount_nullfs:
> 
> BUGS
>      THIS FILE SYSTEM TYPE IS NOT YET FULLY SUPPORTED (READ: IT DOESN'T WORK)
>      AND USING IT MAY, IN FACT, DESTROY DATA ON YOUR SYSTEM.  USE AT
>      YOUR OWN RISK.  BEWARE OF DOG.  SLIPPERY WHEN WET.
> 
> So you can't suggest to use nullfs instead of unionfs, because "is
> well-documented to be broken".

FYI, this is an email I just sent in reply to David Schultz on -hackers@.

%%%

Date: Thu, 10 Mar 2005 12:38:43 +0100
From: Jeremie Le Hen <jeremie@le-hen.org>
To: David Schultz <das@FreeBSD.ORG>
Cc: freebsd-fs@FreeBSD.ORG
Cc: Mikhail Teterin <mi+mx@aldan.algebra.com>
Cc: hackers@FreeBSD.ORG

Hi David,

> Nullfs works better than unionfs.  Unionfs worked well in 4.X.
> Despite numerous minor bugs such as being unable to cope with
> FIFOs, several people have reported using it quite successfully on
> production systems.  However, unionfs no longer works quite as
> well in 5.X or -CURRENT.  There are several reasons for this:
> 
> 1. Nobody seems to have both the time and interest to maintain it.
> 
> 2. Developers can't be expected to prevent regressions in
>    something that's unsupported.
> 
> 3. There are a couple of people who always respond to questions
>    about unionfs with comments along the lines of:
>    ``It's broken, so we won't help you.  Go away and don't tell
>    us if you find any bugs.''
> 
> There's some pretty low-hanging fruit in terms of nits to fix.
> See the PR database if you're interested in helping, and don't let
> anyone scare you away.  ;-)
> 
> > What about the `union' option to regular mounts? Is that safe to use?
> 
> Last I checked, it was very broken, but I'm not sure.

A little time ago, phk@ asked for people to submit regression tests for
virtual filesystem like this [1].  AFAIK, nobody submitted even one test
so far.  This could be a good starting point to have unionfs work
correctly again.  However, I think FreeBSD VFS gurus should first spread
some ideas and clues about tests to do.  I guess indeed there are very
tricky ones that most common mortals wouldn't even suspect.

Regards,

[1] http://lists.freebsd.org/pipermail/freebsd-current/2005-January/045743.html

-- 
Jeremie Le Hen
< jeremie at le-hen dot org >< ttz at chchile dot org >



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20050310120548.GL34822>