From af9b1f9369b13a695d6f83bd540d5f0fb8a56bb6 Mon Sep 17 00:00:00 2001
From: Till JS <tills95@gmail.com>
Date: Tue, 7 Apr 2026 13:19:46 +0200
Subject: [PATCH] fix(mac-mini): make startup.sh idempotent and non-destructive
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

The previous startup.sh checked colima status via `colima status | grep running`
and, if that failed, ran `colima stop --force` unconditionally before starting.
This is destructive: a transient status mis-detection can kill a healthy running
VM, and the subsequent start often hangs because of leftover locks/processes.

Triggered today during the ManaCore→Mana rename: reloading the docker-startup
LaunchAgent ran the script, which falsely concluded colima was down, killed the
running VM, and left 12 zombie limactl processes plus a stale disk lock symlink.
The whole production stack (incl. Forgejo) was offline until manual cleanup.

Changes:
- Use `docker info` as the readiness check instead of `colima status` —
  it directly tests the thing we care about (docker socket reachable)
- Only do cleanup work when we actually need to start; never SIGKILL a
  running VM as a "precaution"
- When we do need to start: reap any zombie limactl/colima processes from
  prior failed runs, and clear the stale disk-in-use lock if no process
  actually holds it
- Verify successful start with `docker info`, not `colima status`

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
---
 scripts/mac-mini/startup.sh | 38 ++++++++++++++++++++++++++++---------
 1 file changed, 29 insertions(+), 9 deletions(-)

diff --git a/scripts/mac-mini/startup.sh b/scripts/mac-mini/startup.sh
index c6353d37e..a6fc468bd 100755
--- a/scripts/mac-mini/startup.sh
+++ b/scripts/mac-mini/startup.sh
@@ -54,15 +54,34 @@ if [ ! -d "/Volumes/ManaData" ]; then
 fi
 
 # ─── Start Colima ───
-if colima status 2>/dev/null | grep -q "running"; then
-    log "Colima already running"
+# Use `docker info` as the source of truth for "is the runtime usable" instead
+# of `colima status`, which can mis-report and trigger a destructive restart.
+if docker info >/dev/null 2>&1; then
+    log "Colima already running (docker reachable)"
 else
+    log "Colima not reachable, preparing fresh start..."
+
+    # Reap zombie colima/limactl processes from previously failed starts.
+    # These hold locks that prevent a clean start. Do NOT touch a running VM.
+    for pat in "colima stop" "limactl stop" "colima daemon" "limactl hostagent" "limactl usernet"; do
+        pids=$(pgrep -f "$pat" || true)
+        if [ -n "$pids" ]; then
+            log "  reaping stale: $pat ($pids)"
+            echo "$pids" | xargs kill -9 2>/dev/null || true
+        fi
+    done
+    sleep 1
+
+    # Clear stale disk lock if no process actually owns it.
+    # The lock is a symlink at /Volumes/ManaData/colima-disk/in_use_by → ~/.colima/_lima/colima
+    # If the symlink exists but no limactl/vz process is running, the lock is stale.
+    LOCK="/Volumes/ManaData/colima-disk/in_use_by"
+    if [ -L "$LOCK" ] && ! pgrep -f "limactl hostagent" >/dev/null 2>&1; then
+        log "  removing stale disk lock: $LOCK"
+        rm -f "$LOCK"
+    fi
+
     log "Starting Colima..."
-
-    # Clear stale process state from hard shutdown (stop only, never delete — delete wipes all images)
-    colima stop --force 2>/dev/null || true
-    sleep 2
-
     colima start \
         --cpu 8 \
         --memory 12 \
@@ -74,8 +93,9 @@ else
         --mount /Volumes/ManaData:w \
         2>&1 | tee -a "$LOG_FILE"
 
-    if ! colima status 2>/dev/null | grep -q "running"; then
-        log "ERROR: Colima failed to start"
+    # Verify with docker info, not colima status (more reliable)
+    if ! docker info >/dev/null 2>&1; then
+        log "ERROR: Colima failed to start (docker not reachable)"
         exit 1
     fi
     log "Colima started successfully"