Replication to localhost creating new dataset in destination dataset from task

[southerner]

Hello all,
I have a recursive snapshot task setup on vol2/pht every day between 10:00-10:05 in which a replication task is setup to send the data everyday at 10:30 to vol3/pht via localhost as the remote machine. My problem is that once the task completes the data's path is now vol3/pht/pht . I have a new dataset of the exact same data as pht. Setting of the task are as follows:
volume/dataset: vol2/pht
remote ZFS: vol3/pht
recursively replicate child's dataset: Checked
delete stale snapshots: checked
...
I am I mis using the replication task?

My need is for a snapshot to send "replicate" the data from vol2/pht to vol3/pht as new data is written to vol2/pht daily.

I have tried to find a script that would utilized ZFS send receive to send incremental streams daily, but had not success there was thread about snycoid but that perl script is nuts.

 

[JustinOtherBobo]

Consider setting up as follows:
remote ZFS: vol3

 

[southerner]

Replication fails without a volume/dataset if the source is volume/dataset. I was able to send the entire pool vol2 to vol3 which included all datasets. I could not just send vol2/pht to vol3.


off topic, I wonder why a ZFS send / recv incremental task has not been created in Freenas.

 

[JustinOtherBobo]

Replication fails without a volume/dataset if the source is volume/dataset. I was able to send the entire pool vol2 to vol3 which included all datasets. I could not just send vol2/pht to vol3.

I am currently using the below dataset structure, snapshot, and replication setup on 11.1-U6 and replication is taking place.

The datasets at LOC1 look as follows when looking at the STORAGE tab:

Code:

pool-loc1
   pool-loc1
        jails
        loc1
            pc1

Tasks are setup as follows:
*snapshot task: pool-loc1/loc1 recursive
*replication task: pool-loc1/loc1 recursive to destination: pool-loc3

In LOC2, STORAGE tab looks as follows:

Code:

pool-loc2
      pool-loc2
           jails
           loc2
              pc1

And tasks are setup as follows:
*snapshot task: pool-loc2/loc2 recursive
*replication task: pool-loc2/loc2 recursive to destination: pool-loc3

The end Results when looking at LOC3, the receiving location:

Code:

pool-loc3
      pool-loc3
          jails
          loc1
                pc1
          loc2
                pc1
          loc3
                some-non-replicated-datasets

However, as I've mentioned in another post, when I clone a snapshot.... say pc2@auto-blah-blah-today and share it out of pool-loc3 to, for example, perform a file recovery at the remote location, the mount-point disappears as soon as a replication receive event starts.

The share still operates and the cloned dataset is still visible under the STORAGE tab.

And as a "new" development, I can confirm that, without having EVER shared out ANYTHING out of the COMPLETE replication, I now have a replication where:

Code:

pool-loc3
      jails
      loc1
          pc1   <-- all files visible at the remote pool's mountpoint
          pc2   <-- all files visible at the remote pool's mountpoint
          pc3   <-- no files are visible at the remote pool's mountpoint;

For pc3 nothing, not even the .zfs folder is visible, it is as if the whole thing doesn't exist. Yet the replication is showing up-to-date and no information is currently getting transmitted.

The snapshots ARE visible in the UI Storage/Snapshot listing.