Installation Manual - 126 spoke-dr-v7 hot standby program plan

The milestone, workstream, phases, batch model, and execution boundary for the spoke-dr-v7 compact hot standby workload cluster.

This chapter records the planning scaffold for spoke-dr-v7, a compact VM-only hot standby workload cluster for spoke-dc-v7.

The intent is to make the DR spoke reproducible and auditable before the first VM is created. spoke-dr-v7 should mirror spoke-dc-v7 logically: same platform structure, GitOps model, operator baseline, storage pattern, policy posture, backup posture, CNPG-backed application data continuity, Open Liberty standby deployment, and failover/failback controls.

The physical topology is intentionally different. spoke-dc-v7 has three VM control-plane nodes and three physical worker nodes. spoke-dr-v7 is planned as a compact three-node VM-only cluster because the current workload is small: one Open Liberty insurance application backed by CloudNativePG.

Governance

Execution Boundary

Live deployment must not begin until issue #465 creates and accepts the new spoke-dr-v7 architecture ADR.

That design gate was accepted on 2026-05-22T19:10:15Z and issue #465 is now closed. The parity baseline gate #466 was completed on 2026-05-22T19:24:59Z. The allocation gate #467 completed at 2026-05-22T19:33:41Z. The next gate is #468.

That ADR must cover:

• compact VM-only workload cluster topology;
• logical parity with spoke-dc-v7;
• CNPG/database-layer replication for the insurance application;
• route and traffic cutover controls;
• split-brain prevention;
• failback and resync principles.

Until that ADR is accepted, do not create VMs, DNS records, HAProxy entries, install artifacts, GitOps overlays, ACM imports, ODF resources, CNPG resources, Open Liberty standby resources, or route cutovers for spoke-dr-v7.

Certification Boundary

Certified in this workstream means internally certified for the CompTech lab:

• every phase has an issue and validation evidence;
• every live change has a rollback or promotion decision;
• install inputs, GitOps overlays, operator state, CNPG replication, app standby, backup, policy, and failover/failback steps are reproducible from documented sources;
• public installation-manual documentation is coherent and batched;
• operational readiness is validated before spoke-dr-v7 is treated as a usable hot standby.

This is not external Red Hat certification, PCI DSS ROC/AOC, QSA, ISO, or bank-auditor attestation.

Non-Goals

• Do not clone spoke-dc-v7 identity, etcd, certificates, kubeconfigs, cluster ID, OAuth secrets, service account signers, RHACS runtime secrets, ACM runtime state, or application secrets.
• Do not let both clusters act as active writers for the database.
• Do not route production traffic to spoke-dr-v7 before the failover drill explicitly allows the traffic shift.
• Do not assume ODF block-level replication for database PVC continuity.
• Do not publish kubeconfigs, kubeadmin passwords, pull secrets, Vault tokens, object-storage credentials, RHACS credentials, HTPasswd values, private keys, PATs, import payloads, or full Secret manifests.

Target State

Candidate VIPs, node IPs, DNS names, VM MACs, disk layout, and final hypervisor placement were assigned by #467. The accepted allocation reserves API/API-int VIP 30.30.230.10, ingress VIP 30.30.230.11, node IPs 30.30.230.13-15, VM names gf-ocp-spoke-dr-v7-master-0..2, MACs 52:54:00:24:00:00-02, and a per-node shape of 32 vCPU, 96 GiB RAM, 300 GiB boot, and 3 x 500 GiB data disks.

Phase Line

Batch Execution Model

The issue line is granular for auditability. Execution should still be batched so each public article reads as one complete operational chapter.

Validation Matrix

Final Done Criteria

spoke-dr-v7 is complete only when all are true:

1. the compact VM-only OpenShift cluster is installed and healthy;
2. hub-dr-v7 imports and manages it cleanly;
3. platform GitOps is synced and healthy;
4. operator, policy, storage, backup, and security parity is complete or explicitly excepted;
5. CNPG replication is validated and promotion is controlled;
6. the Open Liberty standby app is deployed and traffic-isolated;
7. failover and failback/resync evidence is recorded;
8. compliance scans and soak evidence are complete;
9. public installation-manual chapters are comprehensive and current;
10. parent and child issues contain closeout comments with evidence.

Next Gate

The next gate is:

#468 / OP-GF-SPOKEDRV7-4: render install inputs and disconnected artifacts

Batch 0 is complete through allocation. The next work renders install inputs from the accepted allocation and stops before VM creation, which remains gated by #469.