Files
homelab-infra/docs/superpowers/RETIRED.md
T
panxiao81 19cd0a938b
lint / ansible (push) Successful in 3m20s
lint / yaml (push) Successful in 9s
lint / terraform (push) Successful in 32s
lint / yaml (pull_request) Successful in 9s
lint / terraform (pull_request) Successful in 31s
lint / ansible (pull_request) Successful in 4m58s
docs(gitea): 记录精简升级验收策略
2026-09-10 07:40:06 +00:00

1.4 KiB

docs/superpowers — RETIRED 2026-07-28

The plans/ + specs/ workflow is no longer used. Its one piece of work is done and live, so nothing here is pending.

What it covered: the 2026-04-18 migration of shared PostgreSQL to CloudNativePG on OpenEBS ZFS — specs/2026-04-18-postgresql-cnpg-zfs-migration-design.md (design) and plans/2026-04-18-postgresql-cnpg-zfs-migration.md (task-by-task plan).

Outcome: shipped. kubectl get cluster -n shared-db shows shared-postgresql, 1 instance, "Cluster in healthy state", 101 days old, with shared-postgresql-1 Running. The authoritative manifests live in ../../shared-postgresql/ (cloudnativepg-cluster.yaml, shared-postgresql-service.yaml, serviceaccount.yaml), and the dump/restore /cutover/rollback runbook is ../../shared-postgresql/migration.md.

Why retired: the format carried agent-workflow scaffolding ("REQUIRED SUB-SKILL", checkbox task lists) that only ever suited one migration. Design documents now live directly in docs/ — see ../cicd.md — and the split that matters is:

  • CHANGELOG.md — frozen historical snapshot; Git commits and PRs now record changes
  • CLAUDE.md — traps and procedures, for agents
  • docs/*.md — design docs for work not yet built
  • <service>/README.md — how a service actually works

The two files are kept for the migration rationale, which is still the best record of why CNPG and ZFS were chosen. Nothing reads them automatically.