• The first step was to install the HD and notice what device name it had - it is /dev/sdb and it's a 36 GB disk.
• The second step was to run badblocks on it to see if it is suitable for us:

 badblocks -p 5 -s -v -w -o /tmp/ibad /dev/sdb

• The next steps will be to
  • look at the /tmp/ibad file and see how bad it is
  • if it's ok, then partition according to the scheme above
  • print out the partitioning scheme after it is partitioned, with partition sizes for documention purposes
    • maybe tape the paper to the inside of the server
    • as well as email the contents to the sysadmin list)
  • format the partitions with file systems (suggest ext3)
    • use -c option to collect/use badblocks info?
  • use tune2fs to set the -i (interval) and -c (count) options to 0
    • so the system won't do those timeconsuming fsck's on boot
  • mkdir /boot
  • mount /dev/sdbx /boot # whichever partition is appropriate,
    • probably /boot should be at the beginning of the disk
  • copy (using tar or rsync, as root, preserving file attributes, etc) the 9 GB disk to the 36 GB disk root partition
    • (don't copy /tmp, /proc, or /sys (and I'm sure there are a few others that shouldn't be copied))
      • here's trick I use: -IW

mkdir /rootimg
mount -o bind / /rootimg
mount /dev/sd<target> /target
cp -a /rootimg/* /target

• And after that, manual step: we pull the 9 GB drive, put the 36 GB drive in it's place, put another 36 GB drive in the second slot, and see if it will reboot (perhaps after adjusting /etc/fstab on the 36 GB drive).
• Assuming it boots, then we can do most of the above on the 2nd hard drive. We can save the copy of the filesystem contents till after we're live and have decided what to do with all the partitions.