Issue with ZFS replication on 2nd NAS (back up NAS)

I have two Truenas Servers (A and B).
Server B has a ZFS replication task running that pulls snapshots from Server A.
Server B also has a snapshot schedule that takes a snapshot of everything at the beginning of the month. For some reason, this causes the ZFS replication task to no longer work. After Server B takes a snapshot of everything, I get the following error when the ZFS replication tasks runs again:

cannot receive incremental stream: most recent snapshot of tank_1/BackUpTruenas1 does not match incremental source.

Please see attached picture for more info.
It appears that the ZFS replication task on Server B thinks that the latest snapshot is the one created by the snapshot task on Server B at midnight, when in fact the relevant snapshot is the one it pulled from Server A at 1am via ZFS replication task the previous day.

Is there no way to do ZFS replication (pulling snapshots from Server A) and also take Snapshots idependently from the ZFS replication task?

ZFS replication task settings are also attached…


There is a question around what you’re trying to achieve here but giving the snapshot schedule a different naming convention on the backup target should fix the issue.

1 Like

Is a different naming convention really necessary? Shouldn’t the setting “replicate specific snapshots by creation time” (set to 1am) already make sure only the 1am snapshots are being considered?

The reason for taking additional shapshots on Truenas B is so I can roll back in case anything bad happens on Truenas B (accidentally delete something or something gets corrupted etc.)

This video by @Stux may help with seeing the importance of naming and tiered snapshots.

That is not possible.

See the last flowchart in this post.

Those are datasets, not pools, in the flowcharts.

1 Like

i’ve a similar setup. TN Prod ( daily snapshots ) repicate to TN DR. on DR Site i can take separate snapshots as well. you will have to tell the replication task ( dont know the settings out of my head ) that the source destination snapshots dont need to be the same. i can give it a 2nd look when iam back

edit: i rember running into problems when you change anything on the destination, so i left all the replication task to set the destination to “read only”