Replication destroys unaffiliated datasets on destination?

[prionico]

Yeah, please protect other people of this bug. This should be a "blocker" for the next release.

What is the bug ID? I don't find it.

 

[mistermanko]

AFAIK there is no bug report on JIRA and there never was filed any.
If you read along this thread, there is kind of a track record of mishandling the replication task and using it with less knowledge than what would be needed to finish the operation successfully. It was in my case a handling failure, as I've since then, successfully replicated data multiple times. Please read up on each ZFS replication option beforehand and be sure to have understood what the result will be after before executing any set task.

If anyone in this thread can replicate (no pun intended) the "error" please file a bug report. In any other case, use this thread as knowledge database or - more importantly - ask questions and more experienced users here are willing to help - before something goes wrong.

 

[HoneyBadger]

Checking in on this here:

Code:

[2023/02/01 13:31:32] INFO     [Thread-23] [zettarepl.paramiko.replication_task__task_1] Connected (version 2.0, client OpenSSH_8.4p1)
[2023/02/01 13:31:32] INFO     [Thread-23] [zettarepl.paramiko.replication_task__task_1] Authentication (publickey) successful!
[2023/02/01 13:31:32] INFO     [replication_task__task_1] [zettarepl.replication.pre_retention] Pre-retention destroying snapshots: []
[2023/02/01 13:31:32] ERROR    [replication_task__task_1] [zettarepl.replication.run] For task 'task_1' non-recoverable replication error NoIncrementalBaseReplicationError("No incremental base on dataset 'TestStripe/ReallyImportantData' and replication from scratch is not allowed")

SCALE Bluefin refused to obliterate the destination when I directed it to replicate into the root dataset. I'll have to retest on CORE.

Edit: I also get stopped earlier in the process when trying a different scenario (purged destination snaps entirely)

@mistermanko do you remember what version of TrueNAS you were on when this happened?

 

[Whattteva]

My guess for what happened is a combination of (Almost) Full Filesystem Replication (I'm guessing option -R) and option -F on the receiving side, which when combined, will delete any datasets and snapshots that don't currently exist on the source filesystem.

 

[mistermanko]

Checking in on this here:

Code:

[2023/02/01 13:31:32] INFO     [Thread-23] [zettarepl.paramiko.replication_task__task_1] Connected (version 2.0, client OpenSSH_8.4p1)
[2023/02/01 13:31:32] INFO     [Thread-23] [zettarepl.paramiko.replication_task__task_1] Authentication (publickey) successful!
[2023/02/01 13:31:32] INFO     [replication_task__task_1] [zettarepl.replication.pre_retention] Pre-retention destroying snapshots: []
[2023/02/01 13:31:32] ERROR    [replication_task__task_1] [zettarepl.replication.run] For task 'task_1' non-recoverable replication error NoIncrementalBaseReplicationError("No incremental base on dataset 'TestStripe/ReallyImportantData' and replication from scratch is not allowed")

SCALE Bluefin refused to obliterate the destination when I directed it to replicate into the root dataset. I'll have to retest on CORE.

Edit: I also get stopped earlier in the process when trying a different scenario (purged destination snaps entirely)

View attachment 63205

@mistermanko do you remember what version of TrueNAS you were on when this happened?

Just checked my comment history. It must have been CORE 12 U8. Definitely prior CORE 13.