Schema record encoding

[pixie79]

How are differing schema id's handled with the data, or how do you keep schema's in sync between the two clusters?

Just considering that if we write data locally in AVRO with the lead 5 bye header to reference the local schema registry, how can we ensure the same ID match in the central registry? Or is there a way to reencode the header byte to point to the correct central ID during syncing?