Is deduplication pool-wide or dataset-wide only?

I noticed deduplication can be turned off on individual datasets. Does that mean each dataset has its own dedupe table? If multiple datasets with deduplication enabled are on the same pool, will deduplication not occur between them? Wondering if it is pool-wide or contained within datasets. This is not clear in any documentation I’ve seen. My guess is pool-wide, but I don’t want to assume and be wrong.

Thank you!

It’s explained here, specifically the first segment, last paragraph just before the “Benefits” header:

The DDT is pool-wide but the dedup property is per dataset.


Hey @SAK

Deduplication tables are pool-wide; however, please take the time to read through the resource on Deduplication on the TrueNAS Docs hub, especially regarding the Cost/Benefit tradeoffs.