Adding a disk and growing a pool: the step that quietly ends redundancy
Adding a disk is two separate jobs: getting the system to see it, and deciding where it belongs. A mistake in the second can end redundancy for good.
AtlasPVE ·
This entry answers
- proxmox add disk
- add disk to zfs pool
- proxmox expand storage
- proxmox new disk not showing
- zfs pool expansion
Plugging in a new disk looks easy, and usually it is. The risk is not in plugging it in, it is in the decision about where the disk belongs.
First, what is already on the disk
The disk you plugged in may not be new. It can carry an old partition layout, an old pool label, or data left over from another system. Before building anything on it, look at what is already there. Skipping this step is what destroys the data you actually needed, on the very first move.
The decision: separate storage, or grow the pool
Defining the disk as separate storage is the harmless path. The new space comes up as a place that stands on its own and changes the structure of nothing that already exists. For backups, images or less critical machines, this is usually the right answer.
Growing a pool means going inside the existing structure, and that has rules.
The real trap in ZFS
If you add a single disk to a redundant ZFS pool, that disk joins as a separate member with no redundancy of its own. The pool gets bigger, but from that point the entire pool depends on the health of that one disk. If it goes, the pool goes, and the redundancy on the other disks does nothing to help. ZFS warns you at this step, and people routinely force their way past the warning. It is not easy to undo either.
There are two correct ways to grow a redundant pool. Either add a second group of disks in the same shape, or replace the existing disks one at a time with larger ones and wait for each to resilver. The second is slow but needs no extra bays.
How it works in LVM
In LVM, adding the new disk to the group and growing the space is a straightforward job. In return the added disk brings no redundancy: the group is now spread across more disks, and if one of them goes, the data on it goes. Gaining space and gaining durability are not the same thing.
What Atlas does
Atlas shows the attached disks, what is on them and the shape of the pools on one screen. If an operation would lower redundancy, it says so before the operation, not after.
Sources
Proxmox's own documentation. In English, and it has the final word on this subject.