Date: Sat, 26 Jun 2010 12:10:57 +0300 From: Daniel Braniss <danny@cs.huji.ac.il> To: David Wolfskill <david@catwhisker.org> Cc: "Andrey V. Elsukov" <bu7cher@yandex.ru>, freebsd-current@freebsd.org, brucec@freebsd.org Subject: Re: Use of boot0cfg to set boot slice broke between r209459 and r209502 Message-ID: <E1OSRPd-000Kma-ER@kabab.cs.huji.ac.il> In-Reply-To: <20100625180130.GV90096@albert.catwhisker.org> References: <20100625114011.GO90096@albert.catwhisker.org> <201006250837.36220.jhb@freebsd.org> <20100625180130.GV90096@albert.catwhisker.org>
next in thread | previous in thread | raw e-mail | index | archive | help
> > --jr/gb2Ce1GM9KKZD > Content-Type: text/plain; charset=us-ascii > Content-Disposition: inline > Content-Transfer-Encoding: quoted-printable > > On Fri, Jun 25, 2010 at 08:37:36AM -0400, John Baldwin wrote: > > On Friday 25 June 2010 7:40:11 am David Wolfskill wrote: > > > Well, one one of my machines -- I realize that there are some > > > machines for which it's been problematic for a while. And all of > > > the machines I'm using run FreeBSD/i386. > >=20 > > 209469 perhaps? > > Apparently so. > > Here's what I did to test the above assertion: > > * Booted the build machine from slice 4 (usual "head" slice); cloned > that slice to slice 1; booted from slice 1. > > * In a "head" src working directory, I issued > > svn diff -c209469 > > and saw that r209469 merely added 2 lines to usr.sbin/boot0cfg/boot0cfg.c. > > * On the build machine's src working directory, I edited > usr.sbin/boot0cfg/boot0cfg.c to remove the lines in question. > > * Then (as root), I made /usr/src/usr.sbin/boot0cfg/ my current working > directory and issued: > > make && make install > > * I then issued > > boot0cfg -s 4 aacd0 && shutdown -r now > > then watched the serial console. > > * I noticed that the default boot slice -- which had been 1 -- was now > 4. > > * For grins, I then booted slice 4 (head) in single-user mode, mounted > the file systems, then invoked the boot0cfg executable from slice 1 to > switch the default to slice 2, then issued "halt -p". I waited a bit, > then powered the machine back up (WoL can be handy!) noted it was > booting from slice 2, brought it up in single-user mode, then issued > "halt -p" to reduce its power consumption and heat & nouse generation. > > All that said, it looks as if r209469 merely noticed an existing error > condition and tried to do something arguably sensible with it, rather > than merely ignore it. So it would seem that there's a more fundamental > issue at stake, here.... what do you see when you type boot0cfg -v ...? gpart show? then try gpart set -a active -i n aacd0 n will probably be 5. bottom line, the MBR is NOT being updated by boot0cfg danny
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?E1OSRPd-000Kma-ER>