Skip site navigation (1)Skip section navigation (2)
Date:      Tue, 06 Mar 2012 02:21:51 -0600
From:      Doug Barton <dougb@FreeBSD.org>
To:        Adrian Chadd <adrian@freebsd.org>
Cc:        stable@freebsd.org, current@freebsd.org, =?UTF-8?B?eiBXxIVzaWtvd3NraQ==?= <lukasz@wasikowski.net>, Arnaud Lacombe <lacombar@gmail.com>, "Bjoern A. Zeeb" <bzeeb-lists@lists.zabbadoz.net>, Alexander Leidinger <Alexander@leidinger.net>
Subject:   Re: flowtable usable or not
Message-ID:  <4F55C91F.4040503@FreeBSD.org>
In-Reply-To: <CAJ-VmomZ609YvU4XwcW3WiAmBqf51xVa584XVj-1%2B5MF=vzzLg@mail.gmail.com>
References:  <20120221143537.Horde.deyFDZjmRSRPQ52pxBIpnLA@webmail.leidinger.net> <4F4D51CB.2010508@FreeBSD.org> <4F4D5E5D.9040302@FreeBSD.org> <4F4DD288.5060106@FreeBSD.org> <CAHM0Q_O%2BCt6yhRL=B9oxgkL8EgLxZdo7-KFO2C8HqiN1=Kx_bw@mail.gmail.com> <4F4ED889.2070608@FreeBSD.org> <4F500BB9.4040307@FreeBSD.org> <CAHM0Q_OfeB7Kb=pgjGq0uffLJdJROGoCaGz=25Jito-kweAxRQ@mail.gmail.com> <4F5088CA.1090108@FreeBSD.org> <CAHM0Q_MZM6Gn_zPzxz5tLuzPOW=kK9YxqmrLTyitvGfAPhrkbw@mail.gmail.com> <4F510FBD.50008@FreeBSD.org> <CAHM0Q_M7mYV5r_AZEcAxi2wWuQ8u0aXTG-H1YaCLXLUytQd9wA@mail.gmail.com> <4F5117A6.2030003@FreeBSD.org> <CAHM0Q_Mcf=2S_niBf=S6foS7W7pdgKVcOp9n26nMzzc33igp8w@mail.gmail.com> <CAJ-Vmo=3k59ss77AC8VstVfwHruvoH=6tKek7Z1nsL9YWoOXqA@mail.gmail.com> <4F5285CF.3010001@FreeBSD.org> <CAJ-VmomuSTAgtfYUcAepDkCUr5x%2BbVgmdyOhKjBPSSQ5GcxTWw@mail.gmail.com> <4F55B5E3.1080207@FreeBSD.org> <CAJ-VmomZ609YvU4XwcW3WiAmBqf51xVa584XVj-1%2B5MF=vzzLg@mail.gmail.com>

next in thread | previous in thread | raw e-mail | index | archive | help
On 3/6/2012 2:12 AM, Adrian Chadd wrote:
> You haven't been bitten by the storage layer or filesystem hackery
> bits which has caused filesystem corruption. :)

Ummm, I have, actually. I was one of the early adopters of SU+J and
complained loudly when it ate my /var/ for lunch. I also use a lot of
separate slices/partitions, so my system partition isn't getting written
to very often, isn't using SU+J, and almost always comes up clean after
a crash. My layout looks like this:

FreeBSD 1 & 2 are the same:
/ + /usr
/var
/tmp (memory disk)
/usr/local/ (this is the big partition, things like ports WRKDIRPREFIX
and /usr/obj go here)

Then I have separate ext2fs filesystems for /home, /data (cvs, svn,
other big trees). These are accessible from my Linux partition, which is
also where the shared swap partition is.

Using ext2fs for things I really care about (like /home) or things that
would take a long time to reproduce (like cvs and svn trees) has helped
avoid some of the more exciting corruption/data loss events, and
everything on the /usr/local's is either backed up, or trivially
reproducable.

> That said, FFS+SUJ has made recover-from-kernel-panic so much less
> painful. Thankyou Jeffr and others!

It's also made a mess out of snapshots ... The only thing I use SU+J for
is /var and /usr/local (see above).

> What I tend to do is either run current on a VM or organise some
> dedicated -current laptops. And run the bits of -current I'm testing
> on -8 and -9.

Well you get a gold start for actually running it at all, so there you
go. :)


Doug



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?4F55C91F.4040503>