From 27853f7c27d8e1c6bab2e3868c62d2b4258247a1 Mon Sep 17 00:00:00 2001
From: Shane Maynard <snowspeeder@gmail.com>
Date: Thu, 1 Oct 2026 13:34:20 -0400
Subject: [PATCH] fix: restore original services on partial deployment failure

---
 QUALIFICATION.md         | 10 ++++-
 scripts/deploy-unraid.sh | 85 ++++++++++++++++++++++++++++++++++++++++
 2 files changed, 94 insertions(+), 1 deletion(-)
 create mode 100644 scripts/deploy-unraid.sh

diff --git a/QUALIFICATION.md b/QUALIFICATION.md
index 2d4668a..cf18879 100644
--- a/QUALIFICATION.md
+++ b/QUALIFICATION.md
@@ -11,6 +11,8 @@ Target: Unraid `192.168.86.2`, public UI `https://192.168.86.2:9178`.
 - User-facing test login username: `john-doe`. Initial passwords are generated,
   not committed or printed; obtain them from the protected deployment credential
   record or use administrator password reset in the website.
+- Final voice image: `sha256:ea86ac10377c553f32a70d24f46dea1413c949673ed1b9850a4aabbe8574e8d8`;
+  verified `healthy`. Final control HTTP health returned `{"status":"ok"}`.
 - No push performed. The existing dirty NAS checkout was not overwritten; clean
   build source was staged separately. Hope Music and unrelated services untouched.
 
@@ -47,6 +49,10 @@ also passes `model_options.reasoning.enabled: false` and forwards content only.
 - Administrator deletion immediately invalidated session/bootstrap; lifecycle
   removed the account and its named container. Private data retained deliberately.
 - Logout invalidated the token; room grant revocation also covered offline.
+- Actual Azure out-of-band wait-notice and final answer audio received over
+  WebRTC on two independent tracks (measured nonzero peak levels 0.904 / 0.786).
+  Probe used authenticated LiveKit text input with agent voice output and muted
+  synthetic microphone; this verifies synthesis/transport, not physical STT.
 - Actual speech worker registered and final container health passed. Health now
   checks HTTPS, LiveKit, control and PID-bound worker endpoint, not just web shell.
 
@@ -108,6 +114,8 @@ contains no usable key. Keep API origins synchronized for new DNS names.
 Real physical microphone recognition, perceived speech quality, barge-in and
 an installed iPhone against the NAS have **not** been claimed. Self-signed SDK WSS
 trust may require installing a trusted private CA on the iPhone; native HTTP pinning
-is not a guarantee the SDK socket trusts that certificate. No trust-all/plaintext
+is not a guarantee the SDK socket trusts that certificate. The deployed public
+leaf certificate SHA256 was compared with the native HTTP pin and matched exactly
+(`07907abde3bb9c685265909927627d9ea4705d7e7b27056eefe06e10b192af46`). No trust-all/plaintext
 fallback was added. Browser automation's certificate exception is a test setting,
 not production TLS policy. See iOS project handoff for test artifacts.
diff --git a/scripts/deploy-unraid.sh b/scripts/deploy-unraid.sh
new file mode 100644
index 0000000..c568d94
--- /dev/null
+++ b/scripts/deploy-unraid.sh
@@ -0,0 +1,85 @@
+#!/usr/bin/env bash
+set -euo pipefail
+# Run on Unraid after building both images. Never source secret values into logs.
+ROOT=/mnt/user/appdata/hope-voice
+ENV_FILE="$ROOT/runtime.env"
+test -f "$ENV_FILE"
+test -d "$ROOT/accounts"
+test -d "$ROOT/hermes"
+test -f "$ROOT/livekit.yaml"
+test -d /mnt/user/dev/hope-voice-api/certs
+STAMP=$(date -u +%Y%m%dT%H%M%SZ)
+BACKUP="$ROOT/rollback/$STAMP"
+mkdir -p "$BACKUP"
+chmod 700 "$ROOT" "$BACKUP"
+# Capture original immutable IDs before installing state-aware rollback.
+declare -A ORIGINAL
+for name in voice hope-control; do
+  ORIGINAL[$name]=$(docker inspect --format '{{.Id}}' "$name" 2>/dev/null || true)
+done
+rollback() {
+  local code=$?
+  trap - ERR
+  set +e
+  for name in voice hope-control; do
+    current=$(docker inspect --format '{{.Id}}' "$name" 2>/dev/null)
+    # Never remove an original that failed before its rename. Remove only the
+    # replacement occupying the verified fixed name, then restore the original.
+    if [ -n "$current" ] && [ "$current" != "${ORIGINAL[$name]}" ]; then
+      if [ "$(docker inspect --format '{{.Name}}' "$current")" = "/$name" ]; then
+        docker rm -f "$current" >/dev/null
+      fi
+    fi
+    if [ -n "${ORIGINAL[$name]}" ]; then
+      original_name=$(docker inspect --format '{{.Name}}' "${ORIGINAL[$name]}" 2>/dev/null)
+      if [ "$original_name" = "/$name-rollback-$STAMP" ]; then
+        docker rename "${ORIGINAL[$name]}" "$name"
+      fi
+      if [ "$(docker inspect --format '{{.Name}}' "${ORIGINAL[$name]}" 2>/dev/null)" = "/$name" ]; then
+        docker start "${ORIGINAL[$name]}" >/dev/null
+      fi
+    fi
+  done
+  exit "$code"
+}
+trap rollback ERR
+# Configuration receipts contain runtime secrets: root-only, never output them.
+for name in voice hope-control; do
+  if [ -n "${ORIGINAL[$name]}" ]; then
+    docker inspect "$name" > "$BACKUP/$name.json"
+    chmod 600 "$BACKUP/$name.json"
+    test "$(docker inspect --format '{{.Name}}' "$name")" = "/$name"
+    docker stop "$name" >/dev/null
+    docker rename "$name" "$name-rollback-$STAMP"
+  fi
+done
+
+docker run -d --name hope-control --network host --restart unless-stopped \
+  --env-file "$ENV_FILE" \
+  -e CONTROL_DB_PATH=/var/lib/hope/accounts.sqlite3 \
+  -e CONTROL_HOST=127.0.0.1 -e CONTROL_PORT=8092 \
+  -v "$ROOT/accounts:/var/lib/hope" \
+  -v "$ROOT/hermes:$ROOT/hermes" \
+  -v /var/run/docker.sock:/var/run/docker.sock \
+  --log-opt max-size=10m --log-opt max-file=3 hope-voice-control:latest >/dev/null
+for attempt in $(seq 1 30); do
+  if curl -sf http://127.0.0.1:8092/health -o /dev/null; then break; fi
+  if [ "$attempt" = 30 ]; then false; fi
+  sleep 2
+done
+
+docker run -d --name voice --network host --restart unless-stopped \
+  --env-file "$ROOT/voice.env" \
+  -v /mnt/user/dev/hope-voice-api/certs:/etc/voice/certs \
+  -v "$ROOT/livekit.yaml:/etc/livekit.yaml:ro" \
+  --health-cmd='/opt/voice-agent/.venv/bin/python /opt/voice-agent/healthcheck.py' \
+  --health-interval=15s --health-timeout=5s --health-retries=3 --health-start-period=20s \
+  --log-opt max-size=10m --log-opt max-file=3 hope-voice-api-voice:latest >/dev/null
+for attempt in $(seq 1 30); do
+  if [ "$(docker inspect --format '{{.State.Health.Status}}' voice)" = healthy ]; then break; fi
+  if [ "$attempt" = 30 ]; then false; fi
+  sleep 2
+done
+test "$(curl -sk -o /dev/null -w '%{http_code}' -X POST https://127.0.0.1:9178/token)" = 401
+trap - ERR
+printf 'Deployment healthy; rollback containers retained with suffix %s\n' "$STAMP"