Skip to content
Discussion options

You must be logged in to vote

Your interpretation is close, with one important v4 distinction: 0 is a thinning schedule, not “do not take a snapshot.”

For v4 I would use this for the goal you described:

--keep-source=0 --keep-target=1

and I would leave holds enabled for a removable/offline target.

What happens on a successful run:

  1. zfs-autobackup creates a real source snapshot so it has a consistent point-in-time state to send.
  2. It sends that snapshot (incrementally when a common base exists).
  3. v4 creates a source-side ZFS bookmark for the new common point. The bookmark is tagged with the target dataset GUID.
  4. The thinner may then delete the source snapshot because the bookmark can still be used as the base of the next …

Replies: 1 comment 2 replies

Comment options

You must be logged in to vote
2 replies
@Somebodyisnobody
Comment options

@psy0rz
Comment options

Answer selected by Somebodyisnobody
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Category
Q&A
Labels
None yet
3 participants