ZFS, Actually Storage you can finally reason about.

Build your first pool with intent

chain

9 steps, in order — start anywhere

  1. What ZFS Is A system for storing data built around one promise: what you put in comes back exactly—or ZFS tells you, loudly, instead of silently handing you something broken.
  2. The Pool All your disks, combined into one shared pot of storage—and the surprise inside the word: the pool itself keeps no spare copies.
  3. The Vdev The small team of disks that back each other up—where protection actually lives, and therefore where failure actually lives. Lose one whole team, lose the whole pool.
  4. Mirrors Two or more disks holding identical copies of everything—half the space, in exchange for the best speed, the fastest recovery, and the most freedom to change your mind later.
  5. RAIDZ Protection by recovery math instead of full copies—more usable space than mirrors, but a RAIDZ team moves at one disk's pace for scattered small requests. Big libraries, not virtual machines.
  6. Ashift ZFS's recorded guess about a disk's true cell size—made once, when a disk team is created, and permanent forever. The one setting a wrong guess makes unfixable. Use 12.
  7. Datasets A named slice of the pool that looks like a folder but carries its own rules—its own settings, limits, and history. Costs nothing to make. Make many.
  8. Compression Squeezing data used to cost speed to save space. ZFS's method is so fast the trade runs backward—leave it on, and it often makes things faster.
  9. Your First Pool Creating a pool is one command and four judgments—two of them permanent (how the disks are teamed, and the disk cell-size setting), the rest changeable over coffee.