SOLVED Help! 'Replication from scratch is not allowed'.

[NumberSix]

Hi
In an unwise moment, I manually deleted some replication snapshots (if I use the terms correctly - not sure if a 'replication' and a 'snapshot' are the same thing - I think not, not to digress too much here). Now, I have an automated entry under 'Replication Tasks' to create a 'replication' once a week for my 'System' dataset. This week it failed saying
CRITICAL
Replication "System - Scheduled-Backup" failed: No incremental base on dataset 'System' and replication from scratch is not allowed.
However, I cannot find how to fix this, that is to say, how to manually create a replication such that this auto setup will resume working. Here's what I've tried so far:
a) Navigate to Storage/Pools/System dataset/Create snapshot.
b) Navigate to Tasks/Replication Tasks/ Run Now.

Following these steps I changed the frequency of the failed Replication task to 'Hourly' so I can see if it's fixed, rather than waiting a week to find out. It's still failing. Can anyone tell me how to fix the damage I've caused please, and get auto replications running smoothly again? Thank you.

 

[Redcoat]

In your replication task,. you need to check this box:

That will "allow replication from scratch".

 

[NumberSix]

In your replication task,. you need to check this box:

View attachment 48081

That will "allow replication from scratch".

You Sir, are a gentleman and a Scholar. Not to mention a life saver and a star! Thank you so, so much!!

I followed your recommendation and everything has burst into a hive of optimistic sounding activity. The item to be replicated is only a 128meg SSD, but the HDD activity is epic. Fortunately I read the Help icon (?) next to the item you recommended I tick, which suggests the system will first delete every replication that's now out of sync, before starting over with a fresh one. I guess that's what the storm of HDD activity I'm hearing is. And as I write these lines, the activity has concluded.

I have 17 pages of Snapshot results - a tangle of data to my eyes. I don't see anything dated today in there, but I do see refrences to 'System' dating back to May (I set my Snapshots to keep only the last month's worth) so clearly I have other related problems to try to understand. Not for the feint-hearted is it, TrueNAS?!!

Thank you again!

 

[NumberSix]

As a post-script, I see this under Tasks/Periodic Snapshot Tasks

Note where it says 'Pending'. It has always read 'Pending' since I first started TrueNAS back in April. It's supposed to run Snapshots once a week. Does 'Pending' mean it's waiting for next Sunday at midnight or does it mean it's yet to happen for the first time ever?

Thank you.

 

[Redcoat]

does it mean it's yet to happen for the first time ever?

That's what it means.

When it has run successfully (completing its present task) you'll get the lime-green-ground "FINISHED" box.

 

[NumberSix]

Hi
Any idea why my Tasks/Periodic Snapshot Tasks never run, just sit there month after month 'Pending'? The main view is in the image above. Here's one of them in 'Edit' view (they both look identical in edit view, by the way).

The dialogue box isn't that complicated so I can't see any opportunity for me to have done something wrong here, yet the snapshot tasks remain Pending. I am flumoxed! Thanks for your thoughts.

 

[Redcoat]

Did you try to force it to "run now" and seed the process? I think I always did that to check that the config was correct.

 

[NumberSix]

Hi
I see no "Run Now" option in the Periodic Snapshot section.
Are you thinking of the Replication section? - that works since you pointed out which checkbox to tick (above). It's the Periodic Snapshots that have never run.
Thank you.

 

[Redcoat]

Are you thinking of the Replication section?

You are correct - I need to read more carefully.

 

[NumberSix]

Yes indeed.
Unless you have an insight into how it is that these Snapshots never happen, I might open a fresh thread on this problem so it's correctly titled. I admit I've reached this issue as a tangent to this thread's original subject line (which you solved brilliantly! - thank you) .

 

[TN22]

Resurrecting an old thread, but ... why do we have to remove all existing snapshots from the destination before it'll sync with the source? What if I want to have multiple sets of snapshots on my backup volume?

 

[yottabit]

The option referenced in comment #2 is not present in SCALE-22.02.4.

Any idea how to fix this in SCALE? Even from CLI?

Edit: I didn't need the backups in the replication destination, so I just toggle the scratch option on, ran the replication, and then toggled the scratch option off again. Not ideal if I actually wanted to preserve the historical backups, but it at least unblocked me here. The reason this happened was due to a change in one of the source datasets (deleted) being replicated.