install: Disable kube-proxy-replacement by default

We keep kube-proxy-replacement enabled in the external workload test
because that test requires host-reachable services and we want to keep
the same LB configuration for cluster pods and for the external VM.

Related: cilium#15422
Signed-off-by: Thomas Graf <thomas@cilium.io>

cli/.github/in-cluster-test-scripts/external-workloads-install.sh

@@ -8,6 +8,7 @@ cilium install \
   --cluster-name "${CLUSTER_NAME}" \
   --config monitor-aggregation=none \
   --config tunnel=vxlan \
+  --kube-proxy-replacement=strict \
   --native-routing-cidr="${CLUSTER_CIDR}"
 
 # Wait for Cilium status to be ready

cli/internal/cli/cmd/install.go

@@ -67,7 +67,7 @@ cilium install --context kind-cluster1 --cluster-id 1 --cluster-name cluster1
 	cmd.Flags().IntVar(&params.ClusterID, "cluster-id", 0, "Unique cluster identifier for multi-cluster")
 	cmd.Flags().StringVar(&contextName, "context", "", "Kubernetes configuration context")
 	cmd.Flags().StringVar(&params.InheritCA, "inherit-ca", "", "Inherit/import CA from another cluster")
-	cmd.Flags().StringVar(&params.KubeProxyReplacement, "kube-proxy-replacement", "probe", "Enable/disable kube-proxy replacement { disabled | probe | strict }")
+	cmd.Flags().StringVar(&params.KubeProxyReplacement, "kube-proxy-replacement", "disabled", "Enable/disable kube-proxy replacement { disabled | probe | strict }")
 	cmd.Flags().BoolVar(&params.Wait, "wait", true, "Wait for status to report success (no errors)")
 	cmd.Flags().DurationVar(&params.WaitDuration, "wait-duration", 15*time.Minute, "Maximum time to wait for status")
 	cmd.Flags().BoolVar(&params.RestartUnmanagedPods, "restart-unmanaged-pods", true, "Restart pods which are not being managed by Cilium")