Replication issue/question

[onigiri]

Hello there!

I have two identical systems both running freenas 11.3-U1.

Issue:
Replication (built in replication task) via SSH is attempting to send an entire dataset again, despite base and incrementals being present on target.

Situation:
Over the Christmas break, my offsite target machine shutdown due to a power outage and it was not corrected until Monday 1/4/21. Everything came back online, but because one of the datasets was mid-replication, and there was a power shutdown on the target, even though there was a resume token, there was no snapshot on the target to resume. I cleared the resume token and attempt to start sending again. This worked but it is now attempting to send the entire dataset. To add one addition weird element, It is attempting to send 139TB when the dataset is 111TB.

Re-Seeding this data is a massive hassle. Does anyone have any suggestions?

Some screenshots:
Source Snapshots:

Target Snapshots:

Source Replication Task:

Source Dataset:

 

[artlessknave]

the size difference is easy, you have snapshots that are referencing changed data. usually that means you deleted something.
in my experience, once it decides you need to do a full replication...you need to do a full replication. if the checkbox for full sync was checked..it very likely modified the destination and there is no longer a choice. unless you are positive your setup is perfect, you probably want to keep that option off, turning it on only when you want it to run and overwrite.
note that ssh+netcat can make a dramatic difference on the speed.