The disk is full: where the space went and what is safe to delete

When a storage fills up, machines stop, backups fail and panic starts. The worst thing to do at that moment is delete at random. First the space has to be traced, because the three usual culprits carry very different risks when removed.

Snapshots piling up

A snapshot is not a copy; it holds every change made since the moment it was taken. A snapshot left in place for a long time grows by as much data as gets written over it. This is the most common cause of a full storage, and usually the safest cleanup as well: old snapshots you will never return to can go. But look at their dates first, because deleting a snapshot cannot be undone.

Old backups

A backup storage does not clean itself; with no retention rule it grows forever. Deleting here is safe, but the decision is yours: how many versions back do you want to be able to go. Deleting one by one without setting a rule brings the same problem back a few months later.

Orphaned disks

When a machine is deleted its disk does not always leave with it. The disk left behind belongs to no machine but keeps occupying space. These grow quietly because they do not appear in the list. Before deleting, confirm they are truly orphaned: some disks belong to a machine that is merely powered off, and it will want them back when it starts.

Before deleting

On a full storage a hasty decision turns into data loss. The order is this: first see where the space went, then start with the safest, and look again after every step. The lasting way to free space is not a one time cleanup but a retention rule.