One of the features of the dual-write is initial sync where you copy data from the source app (Finance Operation OR DataVerse) to the target app (Finance Operation OR DataVerse) depending on the selection in Master for initial sync option.
This initial sync is the Data Integrator service running behind the scene and copies your data over. You configure the application id for data integrator and add it both apps (Finance Operation OR DataVerse), I have documented it my previous post The Dual Write implementation - Part 2 - understand and action pre-requisites
Master for initial sync can be either Common Data Service (Dataverse) or Finance and Operations apps. For example, If I choose Finance and Operations app in below example where I am syncing Functional Locations then all records will be copied from Finance and operations to Dataverse.
Initial Sync is a full push means if an individual row fails to sync, you cannot resync only failed ones. If the initial synchronization only partially succeeds, a second synchronization runs for all the rows, not just the rows that failed to be synced during the initial synchronization.
1st initial sync Run for 1000 records from FO to CDS à 700 passed and 300 failed
2nd initial sync Run will again run for 1000 records
Do check Considerations for initial Sync from Microsoft Docs
Initial Sync runs against all legal entities configured for dual-write. If you have entered a filter for a specific legal entity in a table map at Finance and Operations app side, as shown below as an example, this will not work for initial sync as it will run against all legal entities configured for dual write under environment details.