Ceph distributed storage

Scope of work

• Sizing and design of a production-grade Ceph cluster.
• Deployment with cephadm and exposure of all three interfaces: block, object, file.
• Integration with OpenStack, Proxmox VE, Kubernetes and external clients.
• RadosGW multi-site replication and RBD mirroring for disaster recovery.
• Performance tuning and recovery of degraded clusters.

Deliverables

• A Ceph cluster in production, correctly sized and documented.
• S3, RBD and CephFS gateways configured and secured.
• Monitoring dashboards and alert thresholds.
• Disk replacement, expansion and upgrade procedures.

Technologies

• Ceph Squid and Reef, cephadm, BlueStore, CRUSH, RadosGW, CephFS, RBD.
• Erasure coding, device classes, NVMe and NVMe-oF.
• Prometheus, Grafana and the Ceph dashboard for observability.

Engagement terms

Ten to thirty days depending on capacity and topology, fixed price.

How the engagement runs

Phase 1 — Sizing

• Analysis of access profiles, capacity and resilience requirements.
• Design of pools, CRUSH rules and dedicated networks.

Phase 2 — Deployment

• Cluster installation and gateway services.
• CephX and TLS hardening, and client onboarding.

Phase 3 — Validation

• Performance and resilience testing: OSD, node and site loss.
• Fine tuning and validation of alert thresholds.

Phase 4 — Operations

• Monitoring and operating procedures put in place.
• Transfer of skills to the storage team.