How to send incremental snapshots to a clone?

[Binarus]

Hello everybody,

how can I send incremental snapshots to a clone?

Scenario (simplified), one pool with name rpool:

zfs create rpool/data
zfs snapshot rpool/data@1
zfs snapshot rpool/data@2
zfs snapshot rpool/data@3
zfs send -R rpool/data@3 | zfs receive -F rpool/backup
zfs destroy rpool/backup@2
zfs clone rpool/backup@1 rpool/clone

This works so far and gives the expected results. However, when I now do

zfs send -I rpool/data@2 rpool/data@3 | zfs receive rpool/clone

I get the error message

cannot receive incremental stream: most recent snapshot of rpool02/server01-temp does not match incremental source

How do I get around this problem?

The goal here is to restore the deleted snapshot rpool/backup@2. Since we can't send old snapshots to a dataset retroactively, the idea is to clone the first snapshot of the dataset in question (rpool/backup), to add all later snapshots to the clone based on that, and finally to promote the clone.

Coming back to the example, I know that I could could leave out the part with the clone and instead delete rpool/backup@3 and incrementally send rpool/data@2 and rpool/data@3 to rpool/backup. However, my situation is more complicated; put short, I don't want to destroy further snapshots in rpool/backup.

So why does sending the incremental snapshot to the clone not work? What I am doing wrong? rpool/backup was a literal copy of rpool/data, and rpool/clone is a literal clone of rpool/backup@1, so the incremental sources should match.

Thanks in advance for any help!

[whoschek]

The 'from' snapshot in the -I syntax should specify a common snapshot, but rpool/data@2 doesn't look like a common snapshot (whereas rpool/data@1 does). If you adjust it like this I guess it would work: zfs send -I rpool/data@1 rpool/data@3 | zfs receive rpool/clone

[IvanVolosyuk]

The problem is the origin snapshot belongs to the original dataset. You can create a send stream on the original dataset from the original dataset from "origin" snapshot, save it, promote the clone temporarily (it will move the origin snapshot to the clone). Then you can receive the steam on clone and after that promote the original dataset to make the clone a clone again. You will have to do that only once, after that both clone and dataset will have common snapshots to deal with.