Issue Encountered with Database Sharding in HAPI FHIR Server

[Ragul Venkatesan]

Hi team,

I wanted to bring to your attention an issue we are facing with our current setup of the HAPI FHIR server. We have implemented multiple PostgreSQL databases for a single HAPI FHIR server using database sharding to address performance issues related to resource searching.

However, we are encountering the following problem:

Initially, we saved 70 resources in Database 1 (DB1).
Subsequently, we saved another 70 new resources in Database 2 (DB2), during which some negative keys were generated.
When attempting to save additional resources in DB2, the system responds with an "ID already exists" error.

Could you please advise on how we might resolve this issue or suggest any best practices for managing resource IDs across sharded databases?

Thanks and regards,

RAGUL V