Isolate node addition e2e from parallel test runs

Run lb-with-node-addition in its own CI step and dedicated Make target, using a Chainsaw label selector instead of the general all test batch. This keeps the cluster-scaling test from mutating node membership while other load-balancer tests are running in parallel.

Also widen the wait windows that were too short for CI: allow up to 10 minutes of 30-second polling for the new MachineDeployment replica to become ready, and extend the IPv6 load balancer readiness gate so backend programming and LB health checks have more time to settle.

Author: komer3

.github/workflows/ci.yml

@@ -119,6 +119,9 @@ jobs:
       - name: Run E2E Tests
         run: devbox run e2e-test
 
+      - name: Run node addition e2e test
+        run: devbox run e2e-test-node-addition
+
       - name: Run Cilium BGP e2e test
         run: devbox run e2e-test-bgp
 

Makefile

@@ -223,7 +223,17 @@ e2e-test:
 	REGION=$(LINODE_REGION) \
 	LINODE_TOKEN=$(LINODE_TOKEN) \
 	LINODE_URL=$(LINODE_URL) \
-	chainsaw test e2e/test --parallel 2 $(E2E_FLAGS)
+	chainsaw test e2e/test --parallel 2 --selector all $(E2E_FLAGS)
+
+.PHONY: e2e-test-node-addition
+e2e-test-node-addition:
+	CLUSTER_NAME=$(CLUSTER_NAME) \
+	MGMT_KUBECONFIG=$(MGMT_KUBECONFIG_PATH) \
+	KUBECONFIG=$(KUBECONFIG_PATH) \
+	REGION=$(LINODE_REGION) \
+	LINODE_TOKEN=$(LINODE_TOKEN) \
+	LINODE_URL=$(LINODE_URL) \
+	chainsaw test e2e/test --selector node-addition $(E2E_FLAGS)
 
 .PHONY: e2e-test-bgp
 e2e-test-bgp:

e2e/test/lb-with-ipv6-backends/chainsaw-test.yaml

@@ -81,19 +81,19 @@ spec:
     - name: Wait for loadbalancer to start serving traffic
       try:
         - script:
-            timeout: 10m
+            timeout: 15m
             content: |
               bash -ce '
               TARGET_IP=$(kubectl get svc svc-test -n "$NAMESPACE" -o json | jq -r '"'"'.status.loadBalancer.ingress[]? | select(.ip != null and (.ip | contains(":") | not)) | .ip'"'"' | head -n1)
               TARGET="http://$TARGET_IP:80"
 
-              for i in {1..24}; do
+              for i in {1..60}; do
                   output=$(curl -s --max-time 8 "$TARGET" | jq -r .podName 2>/dev/null || true)
                   if [[ "$output" == *"test-"* ]]; then
                       echo "loadbalancer ready"
                       exit 0
                   fi
-                  sleep 5
+                  sleep 10
               done
 
               echo "loadbalancer not ready"

e2e/test/lb-with-node-addition/chainsaw-test.yaml

@@ -4,7 +4,7 @@ kind: Test
 metadata:
   name: lb-with-node-addition
   labels:
-    all:
+    node-addition:
 spec:
   namespace: "lb-with-node-addition"
   catch:
@@ -74,7 +74,7 @@ spec:
     - name: Add new node and check nodebalancer gets updated
       try:
         - script:
-            timeout: 10m
+            timeout: 20m
             content: |
               set -euo pipefail
 
@@ -95,7 +95,7 @@ spec:
               required_replicas=$((current_replicas + 1))
               KUBECONFIG=$MGMT_KUBECONFIG kubectl -n "$md_namespace" patch machinedeployment "$md_name" --type='merge' -p "{\"spec\":{\"replicas\":$required_replicas}}"
 
-              for i in {1..18}; do
+              for i in {1..20}; do
                   ready_replicas=$(KUBECONFIG=$MGMT_KUBECONFIG kubectl -n "$md_namespace" get machinedeployment "$md_name" -o=jsonpath='{.status.readyReplicas}')
                   echo "MachineDeployment ready replicas: ${ready_replicas:-0}/$required_replicas"
 
@@ -104,7 +104,7 @@ spec:
                       break
                   fi
 
-                  sleep 10
+                  sleep 30
               done
 
               if [[ ${ready_replicas:-0} -lt $required_replicas ]]; then