On Mon, Jul 11, 2005 at 09:34:52PM +0100, Simon Heywood wrote: > I've had no end of trouble with SILO and RAID-1. On my first few > attempts to install Debian on the Netra, I ran into what I thought was a > bug in silo.c, and hacking around with it produced varying degrees of > success. As far as I can tell, the problem resulted from trying to > install the boot block for a degraded /boot array (only /dev/sdb1 > active). > > Anyway, using the method I described in my previous post, the SILO > installation worked fine. I can boot from disk0:3 or disk1:3, which is > what I wanted all along (the system can completely survive the loss of > one disk). > > S.
I manually got a pair of RAID-1 devices and some LVM configured, finally. I had two small partitions for /boot (sda1/b1 = md0), and two more for the logical volumes (md1). I tricked d-i into using them (I honestly don't even recall now how I did that), and it installed the base system onto my new filesystems. Even SILO was quite happy to use the RAID device; I was able to 'boot disk0' or 'boot disk1' from OBP with identical results. However, the new setup absolutely will *not* complete a boot. Boot process proceeds until it's time to detect and mount the LV's; I get dozens of devfs errors: devfs_mk_dir: invalid argument.<4>devfs_mk_dev: could not append to parent for /disc then the system attempts to proceed, seems mostly okay, then init shuts down my serial console with "/bin/sh: line 1: /dev/pts/0: No such file or directory". I have to manually power cycle the box, as it won't even let me send a break to return to OBP. Ugh. I'll try your suggestion next, probably tomorrow. -- Adam [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
