Snapshot Cannot Be Restored Due to Subsequent Snapshots

March 13, 2026

Sometimes when rolling back an incus container snapshot you may get an error like Error: ... cannot be restored due to subsequent snapshot ... Set zfs.remove_snapshots to override I did. here is the solution

Scenario

We sometimes want to roll a system way back. it should be a simple snapshot restore. Here is the message.

incus snapshot restore CONTAINER 2026-01-18-1737-FRESH
Error: Snapshot "2026-01-18-1737-FRESH" cannot be restored due to subsequent snapshot(s). Set zfs.remove_snapshots to override

If we list the snapshots, we see that we have taken several since so the system is trying to protect us from rolling back too far.

incus snapshot list CONTAINER
+---------------------------------+----------------------+------------+----------+
|              NAME               |       TAKEN AT       | EXPIRES AT | STATEFUL |
+---------------------------------+----------------------+------------+----------+
| 2026-01-18-1737-FRESH           | 2026/01/18 17:37 CST |            | NO       |
+---------------------------------+----------------------+------------+----------+
| 2026-01-18-1738-FRESHER         | 2026/01/18 17:38 CST |            | NO       |
+---------------------------------+----------------------+------------+----------+
| 2026-01-19-0921-DID-THING-1     | 2026/01/19 09:21 CST |            | NO       |
+---------------------------------+----------------------+------------+----------+
| 2026-01-19-1303-DID-THING-2     | 2026/01/19 13:03 CST |            | NO       |
+---------------------------------+----------------------+------------+----------+
| 2026-01-19-1535-DID-THING-3     | 2026/01/19 15:36 CST |            | NO       |
+---------------------------------+----------------------+------------+----------+

Check your storage pools

incus storage list
+---------+--------+-------------+---------+---------+
|  NAME   | DRIVER | DESCRIPTION | USED BY |  STATE  |
+---------+--------+-------------+---------+---------+
| default | zfs    |             | 43      | CREATED |
+---------+--------+-------------+---------+---------+

Since mine is default let's run incus storage show default

incus storage show default
config:
  source: tank/virt/lxd
  volatile.initial_source: tank/virt/lxd
  zfs.pool_name: tank/virt/lxd
description: ""
name: default
driver: zfs
used_by:
...

Closer

zfs list -r tank/virt/lxd
NAME                                                                                            USED  AVAIL  REFER  MOUNTPOINT
tank/virt/lxd                                                                                   211G  1.99T    96K  legacy
tank/virt/lxd/buckets                                                                            96K  1.99T    96K  legacy
tank/virt/lxd/containers                                                                        204G  1.99T    96K  legacy
...
tank/virt/lxd/containers/CONTAINER                                               2.66G  1.99T  3.26G  legacy

So the full path is tank/virt/lxd/containers/CONTAINER and our pool is tank/virt/lxd. We only need the last portion: containers/CONTAINER

To bring it all together, use the location we just found and apply the change just to this CONTAINER's storage, run the following.

incus storage volume set default container/CONTAINER zfs.remove_snapshots=true

Prove it works

incus snapshot restore CONTAINER 2026-01-18-1737-FRESH

The subsequent snapshots are gone; however, so take care.

Another Option

Remember you can also copy from a snapshot. So another option would be powering down the contain and creating a new container from the old snaphot.