RHEL 3.1 kernel update on RAID shows lilo error

[Jonathan Baker-Bates]

I'm running RHEL ES 3.1 i386. The machine has two IDE disks configured as
RAID 1 with md (for all partitions, including /boot), which I did using
diskdruid during the installation. I chose lilo over grub because when I
tried it with grub, the machine would only boot with both disks, while lilo
could from just one when I tested it.

But now I just ran up2date to get the latest kernel errata, and got this:

Installing...
   1:kernel-source          ###########################################
[100%]
   2:kernel                 ###########################################
[100%]
   3:kernel-smp             ###########################################
[100%]
Error installing lilo.conf  The message was:
test install of lilo failed

up2date's log shows:

[Sun May  2 21:33:20 2004] up2date renaming the lilo.conf entry for linux to
linux.bak
[Sun May  2 21:33:20 2004] up2date making a backup copy of /etc/lilo.conf as
/etc/lilo.conf.up2date-1083530000.4200871
[Sun May  2 21:33:20 2004] up2date writing out the new /etc/lilo.conf
[Sun May  2 21:33:20 2004] up2date Running " /sbin/lilo -t -r /"  (lilo test
mode)
[Sun May  2 21:33:20 2004] up2date restoring the backup file
/etc/lilo.conf.up2date-1083530000.4200871 to /etc/lilo.conf
[Sun May  2 21:33:20 2004] up2date Error installing lilo.conf  The message
was: test install of lilo failed

If I try running /sbin/lilo -t -r myself (with -v) I see that there isn't an
"-r" switch (according to man lilo anyway). Is this a bug or something? If I
try running lilo -t -v it gives no errors though.

If I edit lilo.conf by hand and run lilo manually, am I going to be able to
reboot OK? The system is now live so I'd rather not find out the hard way...

Thanks for any help,

Jonathan