I upgraded from Core to Scale and while the server came back online I have critical errors in relation to pool state offline. From what I see it appears I should have resolved my legacy encryption situation before doing the in-place migration. Is there any path to recovering my data or removing the old encryption and importing the pool?
Thanks
Can you not go back to Core?
I am not sure how to go back to Core. The warnings led me to believe this was a one way upgrade.
Do you have your encryption keys? Or a config from core before the upgrade?
If so, you could reinstall core and then re-upload your config
The in-place upgrade to scale changes the boot loader on the boot disk. It will have the core boot environment. so the config can be obtained if you don’t have it, but it can’t boot core, without a new boot loader…
Ie you can install core onto a USB thumb drive if you want.
I have the config and encryption keys available. Once I am back on Core if that’s successful what is the process to remove the legacy encryption?
You can’t.
You can only make the pools available (by unlocking the underlying GELI devices), and then copy the data to another pool (with or without native encryption).
Can you not replace a geli device with a non geli device?
“Yes”
But I don’t think TrueNAS Core will allow it.
It’s “all or nothing” when it comes to legacy encryption (in FreeNAS / TrueNAS Core). You wouldn’t be allowed to add a non-GELI device to a pool that already has GELI encryption in its underlying devices.
In fact, if you want to “replace” a drive in a mirror, for example, you must use the same Encryption Key that the other devices are using.
In other words, it might be too advanced (and risky) for most casual users of TrueNAS.
Instructions to remove geli legacy encryption
You’ll notice they bypass the GUI and use straight ZFS commands.
For a novice, I would feel uneasy for them to try this approach. If they make a mistake (without the aid of the GUI trying to stop them), then they could risk corrupting their pool.
Tread carefully. 
True, but it’s really the only way to get rid of geli in place.
Would be very wise to run a scrub and smart long tests before commencing the process as you will reduce your redundancy by 1 disk during the process.
Looks straightforward enough but as always. “With great power comes great responsibility”
Good find on that article, I’ve not seen that one
Thanks for the assistance. I decided to start fresh on Scale and then copy data over from my backups.
