Devblog

I’ve got one of 1and1 root servers and it just went tits up again and required a reinstall. After going through the Server Re-Image process the server is up and running again – with the following stupid partition setup:

• 1Gb / (on a 120Gb drive!)
• 5Gb /var
• 5Gb /usr
• 109Gb /home

Now unless you run a company creating static HTML files for 100 million small businesses, this is a really stupid setup.

Here’s how you can sort it all out into something more useful. I wanted everything on just one major / partition so I could just forget about it.

• Reboot using the Linux Rescue Mode (from the 1and1 Control Panel)
• The default setup has just enough data on it to fit on the 1GB partition. Copy it all over as follows:
  • mkdir /mnt/root
  • mkdir /mnt/var
  • mkdir /mnt/usr
  • mount -t ext3 /dev/sda1 /mnt/root
  • mount -t xfs /dev/sda5 /mnt/usr
  • mount -t xfs /dev/sda6 /mnt/var
  • mv /mnt/var/* /mnt/root/var
  • mv /mnt/usr/* /mnt/root/usr
• Next: modify fstab for the new setup: use VI to open /mnt/root/etc/fstab and remove the lines containing /var /usr /home
• umount -a
• Next: move your swap to the end of the drive and resize your main / partition to fill the space:
  • Run: “parted”
  • “resize 2 start end” where start and end are near the end of the drive
  • “resize 1 0 near-the-end”
  • If you get errors, try doing the resize in increments
• Reboot and you’re done

All fixed