Should I replace all NAS drives at once or one at a time?

If the goal is to upgrade capacity without losing data, replace drives one at a time, letting the array rebuild fully between each swap. This keeps redundancy intact throughout and gradually migrates the pool onto the new drives.

The one-at-a-time migration

1. Back up the NAS first: a rebuild stresses the remaining drives, and a second failure mid-rebuild can crash the pool.
2. Replace one drive with a larger CMR drive; run Repair; wait for the rebuild to complete.
3. Repeat for each drive, one at a time.
4. Once every drive is the larger size, expand the volume (SHR/RAID 5 release the new space at that point).

Why not all at once

Pulling every drive at once wipes the array: there is no data to rebuild from, so you would create a fresh pool and restore from backup. That is faster if you have a solid backup and do not mind the downtime, but it is not a migration; it is a rebuild-from-scratch. Only do it deliberately, with a verified backup.

A note on drive batches

Replacing drives one at a time over weeks also has a quiet benefit: your drives no longer all come from the same batch and age, reducing the chance of two failing together. Buy CMR NAS drives at least as large as the current ones: the table shows current pricing.