For every unique file, is that file

• stored on a file system that uses an integrated checksum to achieve data integrity?

  Integrated checksums will catch hardware errors, mainly hard drive errors that occur before the drive completely quits.

  Before moving my files to ZFS, I had a Linux file server with mdadm RAID5, and had silent data corruption that mdadm didn’t appear to notice.

  Some file systems that meet this criteria: ZFS, HAMMER2, btrfs

• stored on redundant disks?

  Even with integrated checksums catching read and write errors, hard drives can fail without warning and become paper weights.

  The worst drive failure combination I’ve had is two WD green drives failing within one week of each other. I was quick to replace the drives but came close to losing the 3 disk mirror. Then again, I seem to lose a lot of disks…

• backed up?

  Storing on redundant disks is not backup. Having a backup means the ability to preserve files even through deletion or something like a CryptoLocker virus which encrypts every file on your system and charges for a decrypt. For me, ZFS’s snapshot ability provides this functionality.

• stored off-site?

  I have a drive stored off-site to protect against catastrophic data loss. Some examples: electrical failure (a failing UPS can fry the whole rack), flooding, fire, and theft. This drive is unplugged, and updated monthly.

• encrypted?

  This relates to the ‘theft’ potential above: if someone steals any of my computers, I would like the peace of mind that they cannot recover all of my personal data as well.

The goal of this checklist is to ensure data integrity and safety. Levels of redundancy (disk, machine, even location) allow for many levels of failure before data integrity is affected.