Source pool: zool
Target pool: zima
Replication “zool/mbp - zima” failed: Dataset ‘zool/mbp’ does not have any matching snapshots to replicate..
The source dataset is encrypted. How do I tell the GUI to send the dataset as raw?
That has nothing to do with sending raw streams.
According the middleware, there are no eligible snapshots to replicate. You’ll need to share how you configured the Replication Task. You might be using a naming schema that does not match the Periodic Snapshots.
This is a separate issue. If you want to invoke a raw stream from an encrypted source dataset, you need to enable “Include Dataset Properties” or “Full Filesystem Replication”.
Thank you winnielinnie.
You seem to be suggesting that replication can only work when (after) the automated, scheduled snapshotting service has been enabled for the dataset and has created a snapshot using the expected naming convention.
Is this correct?
Thanks again 
Yes. That is how it was designed in TrueNAS. This is why you connect a Replication Task to a Periodic Snapshot Task. If you don’t do that, then you need to manually specify a naming schema whose pattern matches the snapshots you wish to be replicated.
I allowed TrueNAS to create a scheduled snapshot then tried running the scheduled replication job “manually” i.e. ad hoc and it failed with the below error.
“[EFAULT] Target dataset ‘zima’ does not have matching snapshots but has data (1863654528 bytes used) and replication from scratch is not allowed. Refusing to overwrite existing data.”
I don’t get it. The target should not need a “preexisting” (matching) dataset as far as I know.
truenas_admin has
sudo ALL=(ALL) ALL, NOPASSWD: ALL
on both machines.
Does the target dataset already exist? You want the target to be created by the replication the first time it is run. You should not manually create the destination dataset.
I already see the problem. It appears you selected a pool’s root dataset as the destination. You should add /mpb to the text field for the destination in your replication task so that it reads zima/mpb
I don’t think adding /mbp in the job config is sufficient. I think the remote dataset needs to exist.
I have had success running zfs send jobs manually from the shell. I think I might need to give up on replication using the GUI for today.
That’s not true. In fact, it’s the other way around. You want the replication task to create the destination dataset the first time it runs.