Replication Task - encrypted Pool to unencrypted

[MjSmith]

Hi Guys,

i know this topic is already discussed a lot but i have still problems to understand it.

My Plan (as many before) USB Drive for external Offline Backup.

Source is a encrypted Dataset.
Target is an encrypted Dataset.

Error:
Unable to send encrypted dataset 'DataPool' to existing unencrypted or unrelated dataset 'Backup_USB'.


As far as i read and tested by myself it isnt possible to use the replication task from encrypt to unencrypt right?
Or do i miss something?

Because if i use a unencrypted Dataset as the source everything works fine.

P.S. i would prefer a GUI Solution instead of using console commands to solve this. Call me lazy but i want to keep it simple because i need a stable self running machine were i dont have to study all the time.

best greetings

 

[winnielinnie]

How did you configure the Replication Task?

Can you show a screenshot of all its settings?

 

[MjSmith]

Well sure.
I mean there is nothing special ... i tried also to use full file system as an option or changed read only to ignore without any change.

Is there any solution for this (encrypted)? otherwise i would bite into the apple and rebuild the hole pool by creating a new unencrypted one but i dont understand where should be the problem here? its just a block level job to sync two datasets no matter if the files where encrypted or not ... okay perhaps the os needs an identifiyer to know whats necessary.

i tried also to create an encrypted pool but then it crushed also.

 

[winnielinnie]

This is because you're trying to overwrite a root dataset ("Backup_USB"), which is not possible. (You can't even do this for non-root datasets if there is encryption involved.)

You need to "nest" your destination, so that the first time you run the replication task it will create a new nested child underneath "Backup_USB".

To do this, you can manually type /DataPool or /DataPoolBackup (or whatever you prefer) in the destination field, so that it looks like this:

• Source: DataPool
• Destination: Backup_USB/DataPoolBackup

It's for this reason I use my pseudo-roots method, for more intuitive management. Because, guess what? When you want to "restore" the backup to a newly created pool, you won't be able to maintain the same structure. (You won't be able to "replace" the new pool's root dataset with your backed-up root dataset.)

 

[MjSmith]

Had the same tought already and created a new dataset under Backup_USB and linked that one. Without success.
I try your method now as well.

What made me a little disappointed is that in this article: https://www.truenas.com/docs/scale/...edreplication/replicationwithencryptionscale/

They created a homedir for the admin user and allowed one command without PW.

Is this really necesssary? I mean they wrote it but i dont use homedirs for now. Also a missing?



EDIT: Well it worked ... just with created another (sub?) Dataset within the replication task instead of using an existing one.

EDIT2: and now i experience write and errors on the usb drive yay.

 

[winnielinnie]

Had the same tought already and created a new dataset under Backup_USB and linked that one.

Don't create a new dataset. Let the replication create one for you.


From my post above:

To do this, you can manually type /DataPool or /DataPoolBackup (or whatever you prefer) in the destination field, so that it looks like this:

• Source: DataPool
• Destination: Backup_USB/DataPoolBackup

 

[MjSmith]

thx winnielinnie. Problem is solved with your advise!


But now im challanging new problems ... starting this journey because my mirrored pool gets me errors until it got degraded.
My plan was it to backup everything on a usb drive, get new disk ( im using ssd ) by changing the manufacturer and now the external drives throws also errors...

Not so sure anymore if my problem are really the disks (already changed one) or the controlle or ram itself (32GB ECC).... its so annoying ... this machine cant run 1 month without any problem ...

 

[winnielinnie]

Was your SSD mirror vdev already degraded before you attempted to do this backup to a USB?

Backups and restorations (and data integrity in general) require some pre-planning. Waiting until "something starts to go awry" can lead to permanent dataloss.

USB tends to struggle with prolonged, nonstop I/O. So your initial replication might have triggered or exposed a failing USB drive for all we know.

 

[MjSmith]

it was .... but after i restart the server because i modified it for the new pool the degraded ones return to fully operational. showing me now as it seems a new starting counter.

i had something different in plan ( build up another nas using rsync over vpn ) but ran out of time sadly ... on the good side the data isnt lost. im more afraid about my apps (nextcloud settings and so on...).

failing usb drive means controller of the external case or the drive itself? well i have some drives i could test.... but firstly i want to start memtest to check this thing...