chore: set RAY_gcs_rpc_server_reconnect_timeout_s to reduce wait for …

…gcs to be kill
ray-project · Dec 2, 2024 · c381c31 · c381c31
1 parent 4819287
commit c381c31
Show file tree

Hide file tree

Showing 2 changed files with 5 additions and 2 deletions.
diff --git a/ray-operator/test/e2e/rayservice_ha_test.go b/ray-operator/test/e2e/rayservice_ha_test.go
@@ -258,7 +258,7 @@ func TestRayServiceGCSFaultTolerance(t *testing.T) {
 	// Kill gcs server
 	ExecPodCmd(test, noOpsHeadPod, common.RayHeadContainer, []string{"pkill", "gcs_server"})
 	// wait for head pod not to be ready
-	g.Eventually(HeadPod(test, rayServiceUnderlyingRayCluster), TestTimeoutMedium).Should(WithTransform(sampleyaml.IsPodRunningAndReady, BeFalse()))
+	g.Eventually(HeadPod(test, rayServiceUnderlyingRayCluster), TestTimeoutShort).Should(WithTransform(sampleyaml.IsPodRunningAndReady, BeFalse()))
 
 	go func() {
 		// wait for head pod to be ready

diff --git a/ray-operator/test/e2e/testdata/rayservice.ft.yaml b/ray-operator/test/e2e/testdata/rayservice.ft.yaml
@@ -20,14 +20,17 @@ spec:
             ray_actor_options:
               num_cpus: 1
   rayClusterConfig:
-    rayVersion: '2.9.0'
+    rayVersion: "2.9.0"
     headGroupSpec:
       rayStartParams: {}
       template:
         spec:
           containers:
             - name: ray-head
               image: rayproject/ray:2.9.0
+              env:
+                - name: RAY_gcs_rpc_server_reconnect_timeout_s
+                  value: "20"
               resources:
                 requests:
                   cpu: 300m