Log cause when cluster role creation fails, increase additional timeouts

Signed-off-by: Michael Edgar <medgar@redhat.com>
strimzi · Jan 8, 2024 · 284547c · 284547c
1 parent 281ef4a
commit 284547c
Show file tree

Hide file tree

Showing 3 changed files with 24 additions and 11 deletions.
diff --git a/cluster-operator/src/main/java/io/strimzi/operator/cluster/Main.java b/cluster-operator/src/main/java/io/strimzi/operator/cluster/Main.java
@@ -306,7 +306,8 @@ private static Future<Void> leaderElection(KubernetesClient client, ClusterOpera
             Future.all(futures).onComplete(res -> {
                 if (res.succeeded())    {
                     returnPromise.complete();
-                } else  {
+                } else {
+                    LOGGER.error("Failed to create Cluster Roles.", res.cause());
                     returnPromise.fail("Failed to create Cluster Roles.");
                 }
             });

diff --git a/...rc/test/java/io/strimzi/operator/cluster/operator/assembly/StrimziPodSetControllerIT.java b/...rc/test/java/io/strimzi/operator/cluster/operator/assembly/StrimziPodSetControllerIT.java
@@ -319,11 +319,16 @@ public void testPodCreationDeletionAndRecreation(VertxTestContext context) {
             // Delete the PodSet
             podSetOp().inNamespace(NAMESPACE).withName(podSetName).delete();
 
-            // Check that pod is deleted after pod set deletion
+            /*
+             * Check that pod is deleted after pod set deletion
+             *
+             * The CR finalizer removal does not complete in a consistent amount of time,
+             * therefore the timeout for waiting on the deletion is increased to 1 minute.
+             */
             TestUtils.waitFor(
                     "Wait for Pod to be deleted",
                     100,
-                    10_000,
+                    60_000,
                     () -> client.pods().inNamespace(NAMESPACE).withName(podName).get() == null,
                     () -> context.failNow("Test timed out waiting for pod deletion!"));
 
@@ -655,11 +660,16 @@ public void testNonCascadingDeletion(VertxTestContext context) {
             // Delete the PodSet in non-cascading way
             podSetOp().inNamespace(NAMESPACE).withName(podSetName).withPropagationPolicy(DeletionPropagation.ORPHAN).delete();
 
-            // Check that the PodSet is deleted
+            /*
+             * Check that the PodSet is deleted
+             *
+             * The CR finalizer removal does not complete in a consistent amount of time,
+             * therefore the timeout for waiting on the deletion is increased to 1 minute.
+             */
             TestUtils.waitFor(
                     "Wait for StrimziPodSet deletion",
                     100,
-                    10_000,
+                    60_000,
                     () -> {
                         StrimziPodSet podSet = podSetOp().inNamespace(NAMESPACE).withName(podSetName).get();
                         return podSet == null;

diff --git a/...rc/test/java/io/strimzi/operator/common/operator/resource/StrimziPodSetCrdOperatorIT.java b/...rc/test/java/io/strimzi/operator/common/operator/resource/StrimziPodSetCrdOperatorIT.java
@@ -19,6 +19,7 @@
 import io.strimzi.test.TestUtils;
 import io.vertx.core.Promise;
 import io.vertx.junit5.Checkpoint;
+import io.vertx.junit5.Timeout;
 import io.vertx.junit5.VertxExtension;
 import io.vertx.junit5.VertxTestContext;
 import org.apache.logging.log4j.LogManager;
@@ -28,6 +29,7 @@
 import org.junit.jupiter.api.extension.ExtendWith;
 
 import java.util.Map;
+import java.util.concurrent.TimeUnit;
 import java.util.concurrent.atomic.AtomicReference;
 
 import static org.hamcrest.CoreMatchers.instanceOf;
@@ -190,11 +192,15 @@ public void testUpdateStatus(VertxTestContext context) {
     }
 
     /**
-     * Tests what happens when the resource is deleted while updating the status
+     * Tests what happens when the resource is deleted while updating the status.
      *
-     * @param context   Test context
+     * The CR removal does not consistently complete within the default timeout.
+     * This requires increasing the timeout for completion to 1 minute.
+     *
+     * @param context Test context
      */
     @Test
+    @Timeout(value = 1, timeUnit = TimeUnit.MINUTES)
     public void testUpdateStatusAfterResourceDeletedThrowsKubernetesClientException(VertxTestContext context) {
         String resourceName = getResourceName(RESOURCE_NAME);
         Checkpoint async = context.checkpoint();
@@ -215,10 +221,6 @@ public void testUpdateStatusAfterResourceDeletedThrowsKubernetesClientException(
                     return op.reconcile(Reconciliation.DUMMY_RECONCILIATION, namespace, resourceName, null);
                 })
                 .onComplete(context.succeeding(i -> { }))
-                .compose(i -> {
-                    LOGGER.info("Wait for confirmed deletion");
-                    return op.waitFor(Reconciliation.DUMMY_RECONCILIATION, namespace, resourceName, 100L, 10_000L, (n, ns) -> operator().get(namespace, resourceName) == null);
-                })
                 .compose(i -> {
                     LOGGER.info("Updating resource with new status - should fail");
                     return op.updateStatusAsync(Reconciliation.DUMMY_RECONCILIATION, newStatus.get());