Record the RPF-WP-0015-T03 kubeconfig blocker and round-11 owner reviews.
All checks were successful
CCR intake gate / gate (push) Successful in 1m13s
CI Smoke / host-smoke (push) Successful in 1s
CI Smoke / container-smoke (push) Successful in 2s

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

Assistant: claude-code
Assistant-Model: opus
Assistant-Process: 236506@bnt-lap001
Assistant-Session: 09087018-91e1-48c5-ab2f-acfbb62c92e4
This commit is contained in:
codex 2026-11-03 15:11:34 +01:00
parent cfe3042987
commit 2a1587fc9b

View file

@ -550,3 +550,20 @@ sender lane accepts its receipt as T03
(`tests/test_rpf_wp_0015_e2_receipt.py`). The 2026-10-14/15 blocker was the
KeyCape session at `aal1`. The window needs a fresh password+OTP (`aal2`)
login, and its start must not fall within 20 minutes of 03:17 UTC.
2026-11-03 E2 attempt blocked, then reviewed fix. The founder's preflight
failed at `read node state`. Since RCLUSTER-WP-0019 (2026-10-20) the node
account `tegwick` cannot read the k3s admin kubeconfig, so every `ssh
railiance01 kubectl` call in these scripts fails. The fix (`17fdc27`,
`dad91e8`) adds the opt-in `AUDIT_KUBECONFIG`: the attended admin file from
ops-warden route `kubernetes-railiance01-admin`, set for one command. Cluster
reads, the port-forward and the sender lane's sync, restart and rollback then
run locally; host reads stay on ssh. The round-10 railiance-platform review
requested changes because the lane's writes still went over ssh. That was
fixed before round 11. Round 11 (contract `cd393e26`, base `dad91e8`)
reviews approve from all five owners: audit-core, rapp-postgres,
railiance-cluster, railiance-infra and railiance-platform. These are
owner-scoped sub-agent reviews. The founder posts the hash-bound receipts with
`wp0024-owner-review.py approve` under `AUDIT_KUBECONFIG`, because `approve`
runs the live preflights. Open non-blocking residuals: `created_at` ordering
is a string comparison, and hub `from_agent` is self-declared.