Recursive Dataset Selection Issues With Replication

[suppaduppax]

Last night, I replicated my entire pool (has a periodic recursive snapshot on the root) to a new pool, or so I thought. When I glanced to see if the space used on both pools were in the same ballpark, I noticed that it was off by several terrabytes. After some digging, I saw that the dataset list in the replication task was missing my media dataset. Which is odd because I expected every dataset in the pool to be added to the dataset list in the replication task.

Here is the dataset structure for that media set:
thebank/media
thebank/media/archive

In the replication wizard, when I have recursion on and select thebank/media, it checks both thebank/media AND thebank/media/archive. However, in the source dataset list, it only adds thebank/media/archive. thebank/media does not get added for some reason. Manually selecting each dataset with recursive turned OFF adds both datasets without any issues, but checking all my datasets manually isn't really a great solution. I've included a gif to show what I mean.

So now, I figured i'll just make a new replication task and do thebank/media by itself. But that fails because thebank/media gets created without the periodic snapshots since it was not part of the replication. Now I'm curious if any other datasets were left off the list.

Anyway, I hope someone can help. Thanks!

 

[winnielinnie]

Can you try to create the task without the wizard, using the Advanced page instead?

I noticed it acts wonky in the Wizard page, when I applied the same steps. (Just as you described and displayed).

I never really used the wizard, and I'm not sure if this is a bug or some unexplained "feature" of creating a replication task that doesn't intuitively make sense at first?

 

[suppaduppax]

Would the only dataset in the source list be the root dataset? It doesn't add any child datasets when recursion is on. Does it handle that automatically?

 

[winnielinnie]

Would the only dataset in the source list be the root dataset? It doesn't add any child datasets when recursion is on. Does it handle that automatically?

When using the Advanced page, if you have "Recursive" and/or "Properties" and/or "Full Filesystem" enabled, then simply selecting only the parent dataset implies that everything nested underneath is included.

The replication is based only off of the parent dataset, and since it is recursive, it checks for the same snapshots in the nested children.

If you have,
tank
tank/media
tank/media/downloads
tank/media/streaming
tank/media/isos

Then selecting only tank/media and choosing "Recursive", "Properties", and/or "Full Filesystem", will run the replication on,
tank/media

And it will naturally parse for the same snapshots for,
tank/media/downloads
tank/media/streaming
tank/media/isos

If applicable, that is.

The window might look something like this,
[ ] tank
[X] tank/media
[ ] tank/media/downloads
[ ] tank/media/streaming
[ ] tank/media/isos

 

[suppaduppax]

thanks, i deleted the old pool and i'm gonna try replicating again without the wizard. I'll let you know how it goes.