Deploying an out-of-cluster "production" instance of the Rucio database (using Ceph storage rather than NFS) will both reduce the risk of a significant database failure (as we have observed in the past with postgres data on NFS shares), as well as allowing us to consolidate other database instances in to one single service, thus reducing technical debt and aligning us better with industry best practices.
Deploying an out-of-cluster "production" instance of the Rucio database (using Ceph storage rather than NFS) will both reduce the risk of a significant database failure (as we have observed in the past with postgres data on NFS shares), as well as allowing us to consolidate other database instances in to one single service, thus reducing technical debt and aligning us better with industry best practices.