I have a ZFS filesystem which uses dedup [1]. I'm trying to migrate it to a different pool by using zfs send | zfs receive (as described in Moving a ZFS filesystem from one pool to another for example).
However, the target pool seems to run out of space during the operation. Seems to me that the "dedup" attribute is not copied. I even tried to first manually set dedup to on on the target pool before doing the send / receive, but it still errored out with "out of space".
How can I move the data from one ZFS pool to the other (including all the snapshots that I have!) and ensure that I don't end up using more space on the receiving end?
[1] Yes, I know, there is lots of discussion about "is dedup worth it", but in my case it definitely is because it's "random, often repeated data collected from different storage media" and it allows me to use ~5x less storage.
