If you don’t know what a disk partition is, this piece ain’t for you.
It’s far more difficult for a Linux user to get comfy with Windows than the other way around. That’s because Microsoft Windows works in mysterious ways.
Last night, I decided to install Microsoft Windows XP on my SATA hard-disk, which already had Linux installed on it. The disk had three primary partitions and one extended partition. I emptied one primary partition and deleted it, so as to make space for Windows.
However, when I tried to install Windows, it refused to create a new parition in the free space, claiming that the maximum number of partitions had been reached. Whoa! I’d been under the impression that a maximum of four partitions were supported. If Linux could do it, why not Windows? Anyway, I recreated the deleted partition using Linux, and tried to format the now-existing partition, but our friendly Windows installer refused to oblige, stating that it couldn’t possibly format a partition whose filesystem it could not recognize.
Now here’s the best part: when I decided to forget about installing Windows and deleted the empty partition, the Windows partitioning tool made sure Linux couldn’t boot after that. Somehow, the partition table had been poisoned. I had to boot from a LiveCD of Knoppix to reinstall Grub, and even after that, the 40GB free space on which I had tried to install Windows remained unusable, because the Linux fdisk utility couldn’t recognize it.
I had to take drastic measures. Knowing that many GB of precious data lay stored on my disk, I used fdisk to delete all existing partitions on the disk. Then I recreated all the partitions, specifying the original start and end cylinders of each partition, and rebooted the system.
After a few moments of suspense, the system managed to boot successfully. Whew! Take that, Windows!