Run Now fails if nested datasets are involved

[Basil Hendroff]

Replication is on my radar atm and I have begun moving away from the legacy replication engine of 11.2 to the newer 11.3 replication engine. I believe I've stumbled across a couple of possible anomalies already with the new replication regime (see Replicated dataset is not set to read-only and Local replication concerns). Here's a third possible issue.

Using the replication wizard, if I create a brand new replication for a dataset where there are no pre-existing snapshots...

... and then click RUN NOW, a snapshot is created on the fly and then immediately replicated to the destination.

However, if the source dataset includes nested datasets and no previous snapshots exist, RUN NOW only replicates the parent dataset. Child datasets are not immediately replicated.

Has anyone else noticed this behaviour? I believe it's a bug, but I'd like some confirmation from others before I report it.

The workaround is to edit the periodic snapshot task associated with the replication task and temporarily change the schedule to a custom schedule and a nearby time. This triggers the creation of snapshots for the nested datasets, which then allows the replication to continue.

 

[Basil Hendroff]

Reported in Jira under NAS-107321.

 

[ornias]

I can confirm this, it forced me to create tests for all children, which was annoying.