Date: Mon, 05 Jul 2004 20:37:26 -0600 From: Brad Waite <freebsd@wcubed.net> To: Doug White <dwhite@gumbysoft.com> Cc: freebsd-stable@freebsd.org Subject: Re: 4.10-STABLE boot issues Message-ID: <40EA1066.8030709@wcubed.net> In-Reply-To: <20040619113634.T48022@carver.gumbysoft.com> References: <2398.24.9.172.8.1087316143.squirrel@webmail.wcubed.net> <20040616144028.W16072@carver.gumbysoft.com> <2045.24.9.172.8.1087445008.squirrel@webmail.wcubed.net> <20040619113634.T48022@carver.gumbysoft.com>
next in thread | previous in thread | raw e-mail | index | archive | help
Sorry for the late reply... Work called and I'm just now getting back to *real* important things like my new FreeBSD box. I'd normally trim the quotes, but I'll keep 'em for the context. Doug White wrote: > On Wed, 16 Jun 2004, Brad Waite wrote: > >>>On Tue, 15 Jun 2004, Brad Waite wrote: >>> >>> >>>>I just installed 4.10 (and upgraded to 4.10-STABLE) on a 2 x 9GB SCSI >>>>disk >>>>AMI MegaRAID array (RAID0). I set up slices as I have previously with >>>>30GB+ IDE drives: >>>> >>>> 1GB swap >>>>29GB / >>>> >>>>Fdisk, label and the rest of install went fine, as did the update, but >>>>on >>>>all reboots (even before the update) I got the following error: >>>> >>>>Disk error 0x1 (lba=0x21fc09f) >>>>No /boot/loader >>> >>>From what you got later, it looks like just /boot/loader got corrupted >>>somehow. Go into src/sys/boot/i386 in your 4.10-STABLE tree and do 'make >>>depend && make && make install' as root. That should rebuild & reinstall >>>loader. >> >>It's possible but unlikely that loader is corrupted. I had the same >>problem a few months ago under 4.9. Plus, I already made world, so that >>should have built a fresh version, but it's doing the same thing as the >>stock 4.10 install. >> >> >>>>1. Created a /boot.config containing "0:da(0,a)/kernel" >>> >>>Make sure you remove this when you get loader rebuilt. >> >>Any critical reason for doing this, other than giving me the options found >>in loader? > > > loader arranges for the kernel symbols to be relocated properly, so ps and > friends work again. :-) Otherwise you will get 'no namelist' errors. Hmmm. I don't recall getting those errors when I booted directly to /kernel. > >>>>To further investigate, I copied /boot/loader to / and changed >>>>/boot.config to '0:da(0,a)/loader' and whaddya know? It now runs the >>>>BTX >>>>loader properly. >>>> >>>>So what's going on? Can anyone answer this stumper? >>> >>>Is /boot on /? :-) >> >>Yep. That's what I can't figure out. Is boot2 doing something different >>on a non-root directory because I'm on a RAID controller instead of a >>regular ad or da device? >> >>Oh, I just realized that it can't be a corrupt loader - when I moved it to >>/ it woke up fine. > > > Are you __sure__ /boot is on /dev/da0s1a? You might try booting into > fixit mode from the CDs and running a fsck. I wonder if the directory is > corrupted. > > It obviously works for the rest of the world, and it doesn't look like a > BIOS access problem since it can read the file fine if its on /. Well /boot is *not* on /dev/da0s1a. It's on /dev/amrd0s1a, but boot2 doesn't know how to get at the amr device. At least not as far as I've tested. I'm leery of messing around in /usr/src/sys/boot/i386/boot2/boot2.c, but if someone wants to give me some pointers, I'll try it. There's always the recovery CD...
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?40EA1066.8030709>