Set myself up a test machine today to experiment with local replication before adding an SSD pool and migrating data to it on my primary NAS. Went about recreating some of the conditions of the main NAS to be sure I know expected behaviors and have no issues when I perform the real migrations. As you can see in picture from my test, my HDD pool is unencrypted. There is a dataset named ‘data’ that is also unencrypted. Then a child dataset to that ‘photos’ which is encrypted.
I’ve replicated ‘data’ to the SSD pool ‘fast’ and you can see it’s unencrypted below a pool that is encrypted. Is this an issue? I don’t care that ‘data’ is unencrypted from a safety standpoint, just want to know if it’ll cause actual issues.
Any insight would be greatly appreciated. Thank you
I actually didn’t know that. Thank you for letting me know. Sounds like I should follow that advice then and go with an unencrypted pool. I appreciate the link to the documentation as well.
It can cause a mountpoint to be overlaid with an “empty” folder, since it’s possible for the mount (of the unencrypted dataset) to precede the unlocking of the encrypted parent.
What is ironic is that for some reason TrueNAS SCALE intentionally makes the “Apps” dataset (ix-apps) unencrypted, even if it lives under an encrypted parent.
Remember, with ZFS there are no encrypted or unencrypted “pools”. What’s really going on is if the root dataset is either encrypted or unencrypted.
I appreciate the clarifications. I always appreciate being more properly educated. No doubt my homelab is a graveyard of poor decisions over the years (some of which I’ve “fixed” through rebuilds), but that’s part of the fun in learning.
At least in this case I was smart enough to assemble a system solely for testing so I don’t make these mistakes on my actual in “production” system.
If you’ve got the time, I’d love your opinion on this question I raised in another post.
