fix: auto-clear cpu_manager_state on CPU topology change (#121)

Co-authored-by: Flegma <Flegma@users.noreply.github.com>

Author: Flegma

src/game-server-node/game-server-node.controller.ts

@@ -293,6 +293,30 @@ cat <<-EOF >/etc/rancher/k3s/config.yaml
 	  - "kube-reserved=cpu=1"
 EOF
 
+cat <<-'SCRIPT' >/usr/local/bin/5stack-cpu-state-check.sh
+	#!/bin/bash
+	STATE=/var/lib/kubelet/cpu_manager_state
+	[ ! -f "$STATE" ] && exit 0
+	CACHE="$(dirname "$STATE")/cpu_count"
+	CURRENT=$(nproc)
+	PREVIOUS=$(cat "$CACHE" 2>/dev/null || echo "$CURRENT")
+	if [ "$CURRENT" != "$PREVIOUS" ]; then
+	  echo "CPU count changed from $PREVIOUS to $CURRENT, removing $STATE"
+	  rm -f "$STATE"
+	fi
+	echo "$CURRENT" > "$CACHE"
+SCRIPT
+        chmod +x /usr/local/bin/5stack-cpu-state-check.sh
+
+        mkdir -p /etc/systemd/system/k3s-agent.service.d
+
+cat <<-'DROPIN' >/etc/systemd/system/k3s-agent.service.d/cpu-state-check.conf
+	[Service]
+	ExecStartPre=/usr/local/bin/5stack-cpu-state-check.sh
+DROPIN
+
+        systemctl daemon-reload
+
         rm -f /var/lib/kubelet/cpu_manager_state
 
         systemctl restart k3s-agent