fix: handle Parallels poweroff snapshot restores

2026-03-17 17:33:45 +01:00 · 2026-03-17 04:01:15 +00:00
parent 71a79bdf5c
commit 095a9f6e1d
4 changed files with 196 additions and 27 deletions
--- a/.agents/skills/parallels-discord-roundtrip/SKILL.md
+++ b/.agents/skills/parallels-discord-roundtrip/SKILL.md
@@ -42,10 +42,13 @@ pnpm test:parallels:macos \
 ## Notes

 - Snapshot target: closest to `macOS 26.3.1 fresh`.
+- Snapshot resolver now prefers matching `*-poweroff*` clones when the base hint also matches. That lets the harness reuse disk-only recovery snapshots without passing a longer hint.
+- If Windows/Linux snapshot restore logs show `PET_QUESTION_SNAPSHOT_STATE_INCOMPATIBLE_CPU`, drop the suspended state once, create a `*-poweroff*` replacement snapshot, and rerun. The smoke scripts now auto-start restored power-off snapshots.
 - Harness configures Discord inside the guest; no checked-in token/config.
 - Use the `openclaw` wrapper for guest `message send/read`; `node openclaw.mjs message ...` does not expose the lazy message subcommands the same way.
 - Write `channels.discord.guilds` in one JSON object (`--strict-json`), not dotted `config set channels.discord.guilds.<snowflake>...` paths; numeric snowflakes get treated like array indexes.
 - Avoid `prlctl enter` / expect for long Discord setup scripts; it line-wraps/corrupts long commands. Use `prlctl exec --current-user /bin/sh -lc ...` for the Discord config phase.
+- Full 3-OS sweeps: the shared build lock is safe in parallel, but snapshot restore is still a Parallels bottleneck. Prefer serialized Windows/Linux restore-heavy reruns if the host is already under load.
 - Harness cleanup deletes the temporary Discord smoke messages at exit.
 - Per-phase logs: `/tmp/openclaw-parallels-smoke.*`
 - Machine summary: pass `--json`
--- a/scripts/e2e/parallels-linux-smoke.sh
+++ b/scripts/e2e/parallels-linux-smoke.sh
@@ -14,6 +14,9 @@ INSTALL_VERSION=""
 TARGET_PACKAGE_SPEC=""
 JSON_OUTPUT=0
 KEEP_SERVER=0
+SNAPSHOT_ID=""
+SNAPSHOT_STATE=""
+SNAPSHOT_NAME=""

 MAIN_TGZ_DIR="$(mktemp -d)"
 MAIN_TGZ_PATH=""
@@ -163,7 +166,7 @@ esac
 OPENAI_API_KEY_VALUE="${!OPENAI_API_KEY_ENV:-}"
 [[ -n "$OPENAI_API_KEY_VALUE" ]] || die "$OPENAI_API_KEY_ENV is required"

-resolve_snapshot_id() {
+resolve_snapshot_info() {
  local json hint
  json="$(prlctl snapshot-list "$VM_NAME" --json)"
  hint="$SNAPSHOT_HINT"
@@ -171,28 +174,54 @@ resolve_snapshot_id() {
 import difflib
 import json
 import os
+import re
 import sys

 payload = json.loads(os.environ["SNAPSHOT_JSON"])
 hint = os.environ["SNAPSHOT_HINT"].strip().lower()
 best_id = None
+best_meta = None
 best_score = -1.0
+
+def aliases(name: str) -> list[str]:
+    values = [name]
+    for pattern in (
+        r"^(.*)-poweroff$",
+        r"^(.*)-poweroff-\d{4}-\d{2}-\d{2}$",
+    ):
+        match = re.match(pattern, name)
+        if match:
+            values.append(match.group(1))
+    return values
+
 for snapshot_id, meta in payload.items():
    name = str(meta.get("name", "")).strip()
    lowered = name.lower()
    score = 0.0
-    if lowered == hint:
-        score = 10.0
-    elif hint and hint in lowered:
-        score = 5.0 + len(hint) / max(len(lowered), 1)
-    else:
-        score = difflib.SequenceMatcher(None, hint, lowered).ratio()
+    for alias in aliases(lowered):
+        if alias == hint:
+            score = max(score, 10.0)
+        elif hint and hint in alias:
+            score = max(score, 5.0 + len(hint) / max(len(alias), 1))
+        else:
+            score = max(score, difflib.SequenceMatcher(None, hint, alias).ratio())
+    if str(meta.get("state", "")).lower() == "poweroff":
+        score += 0.5
    if score > best_score:
        best_score = score
        best_id = snapshot_id
+        best_meta = meta
 if not best_id:
    sys.exit("no snapshot matched")
-print(best_id)
+print(
+    "\t".join(
+        [
+            best_id,
+            str(best_meta.get("state", "")).strip(),
+            str(best_meta.get("name", "")).strip(),
+        ]
+    )
+)
 PY
 }

@@ -251,10 +280,42 @@ guest_exec() {
  prlctl exec "$VM_NAME" "$@"
 }

+wait_for_vm_status() {
+  local expected="$1"
+  local deadline status
+  deadline=$((SECONDS + TIMEOUT_SNAPSHOT_S))
+  while (( SECONDS < deadline )); do
+    status="$(prlctl status "$VM_NAME" 2>/dev/null || true)"
+    if [[ "$status" == *" $expected" ]]; then
+      return 0
+    fi
+    sleep 1
+  done
+  return 1
+}
+
+wait_for_guest_ready() {
+  local deadline
+  deadline=$((SECONDS + TIMEOUT_SNAPSHOT_S))
+  while (( SECONDS < deadline )); do
+    if guest_exec /bin/true >/dev/null 2>&1; then
+      return 0
+    fi
+    sleep 2
+  done
+  return 1
+}
+
 restore_snapshot() {
  local snapshot_id="$1"
  say "Restore snapshot $SNAPSHOT_HINT ($snapshot_id)"
  prlctl snapshot-switch "$VM_NAME" --id "$snapshot_id" >/dev/null
+  if [[ "$SNAPSHOT_STATE" == "poweroff" ]]; then
+    wait_for_vm_status "stopped" || die "restored poweroff snapshot did not reach stopped state in $VM_NAME"
+    say "Start restored poweroff snapshot $SNAPSHOT_NAME"
+    prlctl start "$VM_NAME" >/dev/null
+  fi
+  wait_for_guest_ready || die "guest did not become ready in $VM_NAME"
 }

 bootstrap_guest() {
@@ -585,13 +646,16 @@ run_upgrade_lane() {
  UPGRADE_AGENT_STATUS="pass"
 }

-SNAPSHOT_ID="$(resolve_snapshot_id)"
+IFS=$'\t' read -r SNAPSHOT_ID SNAPSHOT_STATE SNAPSHOT_NAME <<<"$(resolve_snapshot_info)"
+[[ -n "$SNAPSHOT_ID" ]] || die "failed to resolve snapshot id"
+[[ -n "$SNAPSHOT_NAME" ]] || SNAPSHOT_NAME="$SNAPSHOT_HINT"
 LATEST_VERSION="$(resolve_latest_version)"
 HOST_IP="$(resolve_host_ip)"
 HOST_PORT="$(resolve_host_port)"

 say "VM: $VM_NAME"
 say "Snapshot hint: $SNAPSHOT_HINT"
+say "Resolved snapshot: $SNAPSHOT_NAME [$SNAPSHOT_STATE]"
 say "Latest npm version: $LATEST_VERSION"
 say "Current head: $(git rev-parse --short HEAD)"
 say "Run logs: $RUN_DIR"
--- a/scripts/e2e/parallels-macos-smoke.sh
+++ b/scripts/e2e/parallels-macos-smoke.sh
@@ -21,6 +21,9 @@ DISCORD_TOKEN_ENV=""
 DISCORD_TOKEN_VALUE=""
 DISCORD_GUILD_ID=""
 DISCORD_CHANNEL_ID=""
+SNAPSHOT_ID=""
+SNAPSHOT_STATE=""
+SNAPSHOT_NAME=""
 GUEST_OPENCLAW_BIN="/opt/homebrew/bin/openclaw"
 GUEST_OPENCLAW_ENTRY="/opt/homebrew/lib/node_modules/openclaw/openclaw.mjs"
 GUEST_NODE_BIN="/opt/homebrew/bin/node"
@@ -291,7 +294,7 @@ cleanup_discord_smoke_messages() {
  discord_delete_message_id_file "$RUN_DIR/upgrade.discord-host-message-id"
 }

-resolve_snapshot_id() {
+resolve_snapshot_info() {
  local json hint
  json="$(prlctl snapshot-list "$VM_NAME" --json)"
  hint="$SNAPSHOT_HINT"
@@ -299,28 +302,54 @@ resolve_snapshot_id() {
 import difflib
 import json
 import os
+import re
 import sys

 payload = json.loads(os.environ["SNAPSHOT_JSON"])
 hint = os.environ["SNAPSHOT_HINT"].strip().lower()
 best_id = None
+best_meta = None
 best_score = -1.0
+
+def aliases(name: str) -> list[str]:
+    values = [name]
+    for pattern in (
+        r"^(.*)-poweroff$",
+        r"^(.*)-poweroff-\d{4}-\d{2}-\d{2}$",
+    ):
+        match = re.match(pattern, name)
+        if match:
+            values.append(match.group(1))
+    return values
+
 for snapshot_id, meta in payload.items():
    name = str(meta.get("name", "")).strip()
    lowered = name.lower()
    score = 0.0
-    if lowered == hint:
-        score = 10.0
-    elif hint and hint in lowered:
-        score = 5.0 + len(hint) / max(len(lowered), 1)
-    else:
-        score = difflib.SequenceMatcher(None, hint, lowered).ratio()
+    for alias in aliases(lowered):
+        if alias == hint:
+            score = max(score, 10.0)
+        elif hint and hint in alias:
+            score = max(score, 5.0 + len(hint) / max(len(alias), 1))
+        else:
+            score = max(score, difflib.SequenceMatcher(None, hint, alias).ratio())
+    if str(meta.get("state", "")).lower() == "poweroff":
+        score += 0.5
    if score > best_score:
        best_score = score
        best_id = snapshot_id
+        best_meta = meta
 if not best_id:
    sys.exit("no snapshot matched")
-print(best_id)
+print(
+    "\t".join(
+        [
+            best_id,
+            str(best_meta.get("state", "")).strip(),
+            str(best_meta.get("name", "")).strip(),
+        ]
+    )
+)
 PY
 }

@@ -377,6 +406,20 @@ resolve_host_port() {
  printf '%s\n' "$HOST_PORT"
 }

+wait_for_vm_status() {
+  local expected="$1"
+  local deadline status
+  deadline=$((SECONDS + TIMEOUT_SNAPSHOT_S))
+  while (( SECONDS < deadline )); do
+    status="$(prlctl status "$VM_NAME" 2>/dev/null || true)"
+    if [[ "$status" == *" $expected" ]]; then
+      return 0
+    fi
+    sleep 1
+  done
+  return 1
+}
+
 wait_for_current_user() {
  local deadline
  deadline=$((SECONDS + TIMEOUT_SNAPSHOT_S))
@@ -458,6 +501,11 @@ restore_snapshot() {
  local snapshot_id="$1"
  say "Restore snapshot $SNAPSHOT_HINT ($snapshot_id)"
  prlctl snapshot-switch "$VM_NAME" --id "$snapshot_id" >/dev/null
+  if [[ "$SNAPSHOT_STATE" == "poweroff" ]]; then
+    wait_for_vm_status "stopped" || die "restored poweroff snapshot did not reach stopped state in $VM_NAME"
+    say "Start restored poweroff snapshot $SNAPSHOT_NAME"
+    prlctl start "$VM_NAME" >/dev/null
+  fi
  wait_for_current_user || die "desktop user did not become ready in $VM_NAME"
 }

@@ -1017,13 +1065,16 @@ FRESH_MAIN_STATUS="skip"
 UPGRADE_STATUS="skip"
 UPGRADE_PRECHECK_STATUS="skip"

-SNAPSHOT_ID="$(resolve_snapshot_id)"
+IFS=$'\t' read -r SNAPSHOT_ID SNAPSHOT_STATE SNAPSHOT_NAME <<<"$(resolve_snapshot_info)"
+[[ -n "$SNAPSHOT_ID" ]] || die "failed to resolve snapshot id"
+[[ -n "$SNAPSHOT_NAME" ]] || SNAPSHOT_NAME="$SNAPSHOT_HINT"
 LATEST_VERSION="$(resolve_latest_version)"
 HOST_IP="$(resolve_host_ip)"
 HOST_PORT="$(resolve_host_port)"

 say "VM: $VM_NAME"
 say "Snapshot hint: $SNAPSHOT_HINT"
+say "Resolved snapshot: $SNAPSHOT_NAME [$SNAPSHOT_STATE]"
 say "Latest npm version: $LATEST_VERSION"
 say "Current head: $(git rev-parse --short HEAD)"
 if discord_smoke_enabled; then
--- a/scripts/e2e/parallels-windows-smoke.sh
+++ b/scripts/e2e/parallels-windows-smoke.sh
@@ -15,6 +15,9 @@ TARGET_PACKAGE_SPEC=""
 JSON_OUTPUT=0
 KEEP_SERVER=0
 CHECK_LATEST_REF=1
+SNAPSHOT_ID=""
+SNAPSHOT_STATE=""
+SNAPSHOT_NAME=""

 MAIN_TGZ_DIR="$(mktemp -d)"
 MAIN_TGZ_PATH=""
@@ -194,7 +197,7 @@ ps_array_literal() {
  printf '@(%s)' "$joined"
 }

-resolve_snapshot_id() {
+resolve_snapshot_info() {
  local json hint
  json="$(prlctl snapshot-list "$VM_NAME" --json)"
  hint="$SNAPSHOT_HINT"
@@ -202,28 +205,54 @@ resolve_snapshot_id() {
 import difflib
 import json
 import os
+import re
 import sys

 payload = json.loads(os.environ["SNAPSHOT_JSON"])
 hint = os.environ["SNAPSHOT_HINT"].strip().lower()
 best_id = None
+best_meta = None
 best_score = -1.0
+
+def aliases(name: str) -> list[str]:
+    values = [name]
+    for pattern in (
+        r"^(.*)-poweroff$",
+        r"^(.*)-poweroff-\d{4}-\d{2}-\d{2}$",
+    ):
+        match = re.match(pattern, name)
+        if match:
+            values.append(match.group(1))
+    return values
+
 for snapshot_id, meta in payload.items():
    name = str(meta.get("name", "")).strip()
    lowered = name.lower()
    score = 0.0
-    if lowered == hint:
-        score = 10.0
-    elif hint and hint in lowered:
-        score = 5.0 + len(hint) / max(len(lowered), 1)
-    else:
-        score = difflib.SequenceMatcher(None, hint, lowered).ratio()
+    for alias in aliases(lowered):
+        if alias == hint:
+            score = max(score, 10.0)
+        elif hint and hint in alias:
+            score = max(score, 5.0 + len(hint) / max(len(alias), 1))
+        else:
+            score = max(score, difflib.SequenceMatcher(None, hint, alias).ratio())
+    if str(meta.get("state", "")).lower() == "poweroff":
+        score += 0.5
    if score > best_score:
        best_score = score
        best_id = snapshot_id
+        best_meta = meta
 if not best_id:
    sys.exit("no snapshot matched")
-print(best_id)
+print(
+    "\t".join(
+        [
+            best_id,
+            str(best_meta.get("state", "")).strip(),
+            str(best_meta.get("name", "")).strip(),
+        ]
+    )
+)
 PY
 }

@@ -338,12 +367,31 @@ restore_snapshot() {
  local snapshot_id="$1"
  say "Restore snapshot $SNAPSHOT_HINT ($snapshot_id)"
  prlctl snapshot-switch "$VM_NAME" --id "$snapshot_id" >/dev/null
+  if [[ "$SNAPSHOT_STATE" == "poweroff" ]]; then
+    wait_for_vm_status "stopped" || die "restored poweroff snapshot did not reach stopped state in $VM_NAME"
+    say "Start restored poweroff snapshot $SNAPSHOT_NAME"
+    prlctl start "$VM_NAME" >/dev/null
+  fi
 }

 verify_windows_user_ready() {
  guest_exec cmd.exe /d /s /c "echo ready"
 }

+wait_for_vm_status() {
+  local expected="$1"
+  local deadline status
+  deadline=$((SECONDS + TIMEOUT_SNAPSHOT_S))
+  while (( SECONDS < deadline )); do
+    status="$(prlctl status "$VM_NAME" 2>/dev/null || true)"
+    if [[ "$status" == *" $expected" ]]; then
+      return 0
+    fi
+    sleep 1
+  done
+  return 1
+}
+
 wait_for_guest_ready() {
  local deadline
  deadline=$((SECONDS + TIMEOUT_SNAPSHOT_S))
@@ -830,13 +878,16 @@ run_upgrade_lane() {
  UPGRADE_AGENT_STATUS="pass"
 }

-SNAPSHOT_ID="$(resolve_snapshot_id)"
+IFS=$'\t' read -r SNAPSHOT_ID SNAPSHOT_STATE SNAPSHOT_NAME <<<"$(resolve_snapshot_info)"
+[[ -n "$SNAPSHOT_ID" ]] || die "failed to resolve snapshot id"
+[[ -n "$SNAPSHOT_NAME" ]] || SNAPSHOT_NAME="$SNAPSHOT_HINT"
 LATEST_VERSION="$(resolve_latest_version)"
 HOST_IP="$(resolve_host_ip)"
 HOST_PORT="$(resolve_host_port)"

 say "VM: $VM_NAME"
 say "Snapshot hint: $SNAPSHOT_HINT"
+say "Resolved snapshot: $SNAPSHOT_NAME [$SNAPSHOT_STATE]"
 say "Latest npm version: $LATEST_VERSION"
 say "Current head: $(git rev-parse --short HEAD)"
 say "Run logs: $RUN_DIR"