pdurlej/platform

Fork 0

Phase 8 monthly restore-test stale 26+ days #45

New issue

Closed

opened 2026-05-03 23:30:22 +02:00 by claude · 2 comments

claude commented

2026-05-03 23:30:22 +02:00

Collaborator

Migrated from state/L3/OPEN_LOOPS.md per ADR 0001 + AGENTS.md (Forgejo Issues as memory layer for follow-ups).

Source

state/L3/OPEN_LOOPS.md Backup/DR validation gate section + STATE_OF_PLATFORM v2 §5 risk/runtime HIGH.

What

Phase 8 monthly restore-test (originally af8855ae Feb 18 plan) — required before 'production declare'. hp-restore-smoke.timer last ran 2026-04-07, 26+ days stale.

Reclassification (per operator amendment 14, 2026-05-03)

NOT immediate Phase 02 blocker. Critical pre-cutover gate. Blocks production declare / final RS2000 replacement. Local restore rehearsal may happen earlier to reduce cutover risk.

Default action (from STATE_OF_PLATFORM v2 Owner Action Board)

Resume monthly cadence; first run scheduled within 7 days. Honcho-postgres-specific quarterly drill on top.

Owner

piotr — schedule the first run; verify timer re-armed.

Acceptance criteria

First restore-test run completes successfully within 7 days
hp-restore-smoke.timer shows recent run (systemctl --user status hp-restore-smoke.timer)
Honcho-postgres-specific quarterly drill scheduled in calendar
Issue closed when both monthly cadence + quarterly Honcho drill verified active

Migrated from `state/L3/OPEN_LOOPS.md` per ADR 0001 + AGENTS.md (Forgejo Issues as memory layer for follow-ups). ## Source `state/L3/OPEN_LOOPS.md` Backup/DR validation gate section + STATE_OF_PLATFORM v2 §5 risk/runtime HIGH. ## What Phase 8 monthly restore-test (originally af8855ae Feb 18 plan) — required before 'production declare'. `hp-restore-smoke.timer` last ran 2026-04-07, **26+ days stale**. ## Reclassification (per operator amendment 14, 2026-05-03) NOT immediate Phase 02 blocker. Critical pre-cutover gate. Blocks production declare / final RS2000 replacement. Local restore rehearsal may happen earlier to reduce cutover risk. ## Default action (from STATE_OF_PLATFORM v2 Owner Action Board) Resume monthly cadence; first run scheduled within 7 days. Honcho-postgres-specific quarterly drill on top. ## Owner piotr — schedule the first run; verify timer re-armed. ## Acceptance criteria - [ ] First restore-test run completes successfully within 7 days - [ ] `hp-restore-smoke.timer` shows recent run (`systemctl --user status hp-restore-smoke.timer`) - [ ] Honcho-postgres-specific quarterly drill scheduled in calendar - [ ] Issue closed when both monthly cadence + quarterly Honcho drill verified active

claude added the

labels

2026-05-03 23:30:47 +02:00

pdurlej referenced this issue

2026-05-05 02:03:47 +02:00

ops(security): Codex SSH key delivery via ssh-agent with TTL — keep key off disk (Phase 1 Lite) #73

pdurlej referenced this issue

2026-05-05 02:08:26 +02:00

arch(governance): kontrakt + granica między pdurlej/platform a pdurlej/iskra-openclaw — co gdzie mieszka #74

codex referenced this issue

2026-05-05 03:18:25 +02:00

feat(agent-access): add codex OpenClaw ssh-agent TTL wrapper #77

claude referenced this issue

2026-05-06 00:26:05 +02:00

feat(modules/honcho-postgres): upgrade manifest to v2 cataloging #99

claude referenced this issue

2026-05-06 00:26:06 +02:00

feat(modules/synapse): upgrade manifest to v2 cataloging #104

codex referenced this issue

2026-05-09 01:01:23 +02:00

docs(module): upgrade honcho-postgres catalog to v2 #127

codex referenced this issue

2026-05-17 18:27:00 +02:00

docs(cutover): plan post-soak legacy cleanup #340

codex added this to the 02 - DR and restore confidence milestone

2026-05-19 08:36:20 +02:00

codex referenced this issue

2026-05-24 16:21:40 +02:00

docs(dr): start W3 restore confidence #430

codex commented

2026-05-24 16:31:58 +02:00

Collaborator

Codex W3b restore smoke — 2026-05-24 16:28 CEST

Role: executor
Status: green, metadata-only evidence PR opened: #431

Evidence:

Operator gate received: w3-restore-smoke-approved.
hp-restore-smoke.service started manually and exited 0/SUCCESS.
Duration: 6s.
Backup used: /opt/vps-home-platform-infra/backups/20260524-120007-critical.
Restore image: postgres:16.12-alpine.
Unhealthy containers after run: 0.
Disposable restore-test-postgres container count after run: 0.
Timer remains active; next scheduled run: 2026-06-02 03:45 CEST.

Interpretation: the monthly Forgejo restore-smoke path is no longer stale. I am not closing this issue from the API because the original acceptance also mentions Honcho quarterly scheduling; #238 remains the broader W3 carrier.

## Codex W3b restore smoke — 2026-05-24 16:28 CEST **Role:** executor **Status:** green, metadata-only evidence PR opened: #431 Evidence: - Operator gate received: `w3-restore-smoke-approved`. - `hp-restore-smoke.service` started manually and exited `0/SUCCESS`. - Duration: 6s. - Backup used: `/opt/vps-home-platform-infra/backups/20260524-120007-critical`. - Restore image: `postgres:16.12-alpine`. - Unhealthy containers after run: `0`. - Disposable `restore-test-postgres` container count after run: `0`. - Timer remains active; next scheduled run: 2026-06-02 03:45 CEST. Interpretation: the monthly Forgejo restore-smoke path is no longer stale. I am not closing this issue from the API because the original acceptance also mentions Honcho quarterly scheduling; #238 remains the broader W3 carrier.

claude referenced this issue

2026-05-24 18:13:47 +02:00

docs(dr): record Honcho partial restore #432

codex referenced this issue

2026-05-24 18:18:23 +02:00

dr(w3d): full sandbox DR drill — operator-gated #433

codex commented

2026-05-24 18:18:46 +02:00

Collaborator

W3a/b/c restore-confidence evidence is now current and accepted as the immediate gate per PR #432 comment #9292.

Evidence package:

#430: W3 preflight reconciled current backup/restore timer state.
#431: W3b reran hp-restore-smoke.service; exit 0/SUCCESS, disposable Forgejo PostgreSQL restore cleaned up, zero unhealthy containers.
#432: W3c restored Honcho pgvector backup into an isolated disposable target; metadata-only schema/count/dimension checks passed, zero unhealthy containers.

The stale monthly restore-test risk described here is no longer accurate: the restore smoke was rerun successfully on 2026-05-24, and the timer remains active for the next scheduled run.

The deeper full sandbox DR drill is still tracked separately in #433 and is required before irreversible Class A/B/D legacy cleanup or broad W8 module upgrade waves. Closing this stale cutover-gate issue now.

W3a/b/c restore-confidence evidence is now current and accepted as the immediate gate per PR #432 comment #9292. Evidence package: - #430: W3 preflight reconciled current backup/restore timer state. - #431: W3b reran `hp-restore-smoke.service`; exit `0/SUCCESS`, disposable Forgejo PostgreSQL restore cleaned up, zero unhealthy containers. - #432: W3c restored Honcho pgvector backup into an isolated disposable target; metadata-only schema/count/dimension checks passed, zero unhealthy containers. The stale monthly restore-test risk described here is no longer accurate: the restore smoke was rerun successfully on 2026-05-24, and the timer remains active for the next scheduled run. The deeper full sandbox DR drill is still tracked separately in #433 and is required before irreversible Class A/B/D legacy cleanup or broad W8 module upgrade waves. Closing this stale cutover-gate issue now.

codex closed this issue

2026-05-24 18:18:47 +02:00