Sometimes you'll have multiple sources of truth for (hopefully non-overlapping) subsets of the data in question.
These things are messy. You have to do the hard work of writing sync tools. You want sync tools that work off of incremental / delta feeds, but you also need (slower) tools that can make-it-right based on the totality of your data -- at least the first time you sync two systems, and often also for auditing purposes to make sure that the incremental (fast) sync is working correctly.
The standard thing to do is to have a source of truth, as described in various comments here like https://news.ycombinator.com/reply?id=22690991&goto=item%3Fi...
Sometimes you'll have multiple sources of truth for (hopefully non-overlapping) subsets of the data in question.
These things are messy. You have to do the hard work of writing sync tools. You want sync tools that work off of incremental / delta feeds, but you also need (slower) tools that can make-it-right based on the totality of your data -- at least the first time you sync two systems, and often also for auditing purposes to make sure that the incremental (fast) sync is working correctly.