Noobsauce is right, the datasets are like a virtual partition of your pool, it's all or nothing.
I think you misunderstood my earlier post. I was suggesting you run multiple, successive zdb -e -bcsvL tank
So run it, wait for it to finish, run it again, and again, don't try importing unless it comes back without an error. Each successive run should hopefully make more progress until it finishes without an error.
I think you misunderstood my earlier post. I was suggesting you run multiple, successive zdb -e -bcsvL tank
So run it, wait for it to finish, run it again, and again, don't try importing unless it comes back without an error. Each successive run should hopefully make more progress until it finishes without an error.