Hi,

As part of some exploratory work on health-aware, DNS-driven Load Balancing for IBM Storage Ceph Object (RGW), I wrote a 5-part blog series to share the findings:

Part 1: Building Health-Aware Load Balancing for IBM Storage Ceph Object Gateway with Consul

Per-node HAProxy "concentrators" fronting local RGWs, registered and health-checked in Consul.

Part 2: Seamless DNS Routing with CoreDNS and Virtual-Host Style Bucket Access

CoreDNS on port 53 rewrites *.s3.cephlab.com to Consul service domain, supports virtual-host buckets, and plays nicely with enterprise DNS (stub delegation).

Part3. Consul-Powered Global Load Balancing for IBM Storage Ceph Object Gateway (RGW).

In Parts three and four, we combine IBM Storage Ceph Object Multisite (active/active data replication between Madrid and Paris Data Centers) with a Consul global load-balancing control plane, ensuring that the single S3 endpoint always routes to the closest healthy ingress and fails over across sites when needed, without requiring any client URL or SDK changes. 

Part 4. Consul-Powered GLB for Active/Active IBM Storage Ceph Object Multisite Replication

Turn two or more IBM Storage Ceph Object sites into one global S3 endpoint.
We paired IBM Storage Ceph Object Multisite Replication (active/active for data) with Consul to build a global control plane. Each site routes locally when healthy and fails over to other sites only when necessary. Clean architecture, small blast radius, fast convergence.

Part 5. Consul-Powered Active/Passive Disaster Recovery for IBM Storage Ceph Object Multisite Replication

Thanks!