I am thinking of changing my current homeserver - HP Microserver 8 to newer hardware and turning this box into a TrueNAS type machine just for storage and backups.
I currently have 4No x 6Tb with a 500Gb SSD that runs the Windows OS at the moment.
Can i over time, change these disks to be bigger storage. Eg If i replace the current drives with 10Tb drives in stages over say a year (Cant afford to do it in one go!). Can I set the storage to be bigger once i complete the swap ? Without losing the data currently on the drives ?
Or will i have to rebuild it ?
Thanks
If you have 4 drives and build 2 mirror vdevs out of them (raid10). You get ~12T of usable space.
You can replace 2 drives in a mirror vdev one by one with e.g. 10T drives, after replacing both you’ll have, 16T of space instead of 12 before.
If you build a RaidZ1 of 4 disks. You’ll have 18T of usable space, to grow the space you’ll need to replace all 4 of them one at a time.
Filesystem is online while this is happening, but slower.
If you’d like more flexibility, you can try Unraid (Linux based), or you can try OpenMediaVault or Rockstor, or you could try BTRFS by yourself if you have some Linux experience and/or patience to learn how. They all allow you to add disks/remove disks and change raid levels and rebalance / reencode data to more disks or fewer arbitrarily-ish, without going offline while data moves around and without rebuilding the filesystem from scratch.
Can i over time, change these disks to be bigger storage.
The last time I evaluated it, ZFS was rather inflexible about reshaping arrays. The enterprise pattern is to migrate to new storage, and that’s reflected in ZFS’s robust send features.
Autoexpand? resizes when vdev completely on larger drives?
Single vdev only changes when all drives don, a pool of mirrors can change when each mirror/vedev is changed?