File systems, partitions, and raids - lessons learned

Here are a few basic rules I’ve learned about file systems, partitions, and raid devices:

  1. grub error with xfs
  1. It is unusual to boot off of a raid device, and most systems will not do it
  2. Never use reiserfs - he murdered his wife!
  3. Don’t use ext3 for large filesystems with large files (for example, 500gb partition dedicated to a/v media)
  4. There are way more raid levels than necessary. Either go fast, redundant, or spend the $$ for both. That’s three.
  5. Don’t ever try to boot off of xfs with grub
  6. Don’t even try to use jfs, whose support has been dropped by many a distro now
  7. Installations happen faster if you are also watching dragonLance
  8. Always make two swap partitions
  9. Make sure you know how to optimize and properly initialize whatever you do use. For example, make sure it is formatted with a high number of inodes if you expect many files. Or, if using raid, align the stripe sizes to proper block sizes (like 64).
  10. Switch off filesystem’s logging of access times during mount. The main function of access time logging is to make your filesystem slower. Seriously. Nobody ever uses that feature. ‘mount -o noatime,nodiratime’
  11. I know there should only be ten rules, but seriously, practice safety and use truecrypt

2 Responses to “File systems, partitions, and raids - lessons learned”

  1. FloorFLUX Says:

    What’s the purpose of two swap partitions? Can’t two seperate installs share a single one?

  2. BloodKillerDeathMan Says:

    Superstition man. Pure superstition. Also, if one somehow screws up, you don’t want to be resizing partitions, formatting new drives, etc. Like the karate kid: “swapoff”, “swapon”.

Leave a Reply