On Sun, Jan 03, 2010 at 12:03:00PM -0500, Kent Watsen wrote:
[upgrading a system that uses RAIDframe]
> What do other people do?
I'm now running a GENERIC kernel with an mpi RAID card, but I've used
RAIDframe for several years. I had a disk die on me, and RAIDframe worked
great. The disk suddenly seemed to disappear from the system, so any
command sent to it resulted in a very noticeable timeout, during which
process scheduling didn't function. After some time, RAIDframe marked the
disk as failed and didn't issue any more I/O's to it and the system worked
normally after that. (Note that this is one of the things you should take
into account when using a watchdog: don't set its timeout too low.)
Anyway, what I did was create several normal, non-RAID, partitions
and one large RAID partition on each disk. On the normal partitions of each
disk, I performed a normal OpenBSD install. Using that installation, I
created two RAIDframe enabled kernels: one with and one without RAID
autoconfig. I can then reboot into the non-autoconfig RAIDframe enabled
kernel, manually configure the RAIDframe array, and then upgrade the
OpenBSD installation within. Once happy, I can set the boot.conf of the two
disks to the autoconfig RAIDframe enabled kernel, which will then load from
one of the normal partitions, but once started will use the root filesystem
in the RAID partition.
--
Jurjen Oskam
Savage's Law of Expediency:
You want it bad, you'll get it bad.
