Promote ContinueReconciliationOnManualRollingUpdateFailure feature …

…gate to beta (#10524) Signed-off-by: Jakub Scholz <www@scholzj.com>
strimzi · Sep 2, 2024 · 36f1109 · 36f1109
1 parent e7e8b11
commit 36f1109
Show file tree

Hide file tree

Showing 15 changed files with 52 additions and 72 deletions.
diff --git a/.azure/templates/jobs/system-tests/feature_gates_regression_jobs.yaml b/.azure/templates/jobs/system-tests/feature_gates_regression_jobs.yaml
@@ -5,7 +5,7 @@ jobs:
       display_name: 'feature-gates-regression-bundle I. - kafka + oauth'
       profile: 'azp_kafka_oauth'
       cluster_operator_install_type: 'bundle'
-      strimzi_feature_gates: '+ContinueReconciliationOnManualRollingUpdateFailure'
+      strimzi_feature_gates: '-ContinueReconciliationOnManualRollingUpdateFailure'
       strimzi_use_node_pools_in_tests: "false"
       timeout: 360
       releaseVersion: '${{ parameters.releaseVersion }}'
@@ -17,7 +17,7 @@ jobs:
       display_name: 'feature-gates-regression-bundle II. - security'
       profile: 'azp_security'
       cluster_operator_install_type: 'bundle'
-      strimzi_feature_gates: '+ContinueReconciliationOnManualRollingUpdateFailure'
+      strimzi_feature_gates: '-ContinueReconciliationOnManualRollingUpdateFailure'
       strimzi_use_node_pools_in_tests: "false"
       timeout: 360
       releaseVersion: '${{ parameters.releaseVersion }}'
@@ -29,7 +29,7 @@ jobs:
       display_name: 'feature-gates-regression-bundle III. - dynconfig + tracing + watcher'
       profile: 'azp_dynconfig_listeners_tracing_watcher'
       cluster_operator_install_type: 'bundle'
-      strimzi_feature_gates: '+ContinueReconciliationOnManualRollingUpdateFailure'
+      strimzi_feature_gates: '-ContinueReconciliationOnManualRollingUpdateFailure'
       strimzi_use_node_pools_in_tests: "false"
       timeout: 360
       releaseVersion: '${{ parameters.releaseVersion }}'
@@ -41,7 +41,7 @@ jobs:
       display_name: 'feature-gates-regression-bundle IV. - operators'
       profile: 'azp_operators'
       cluster_operator_install_type: 'bundle'
-      strimzi_feature_gates: '+ContinueReconciliationOnManualRollingUpdateFailure'
+      strimzi_feature_gates: '-ContinueReconciliationOnManualRollingUpdateFailure'
       strimzi_use_node_pools_in_tests: "false"
       timeout: 360
       releaseVersion: '${{ parameters.releaseVersion }}'
@@ -53,7 +53,7 @@ jobs:
       display_name: 'feature-gates-regression-bundle V. - rollingupdate'
       profile: 'azp_rolling_update_bridge'
       cluster_operator_install_type: 'bundle'
-      strimzi_feature_gates: '+ContinueReconciliationOnManualRollingUpdateFailure'
+      strimzi_feature_gates: '-ContinueReconciliationOnManualRollingUpdateFailure'
       strimzi_use_node_pools_in_tests: "false"
       timeout: 360
       releaseVersion: '${{ parameters.releaseVersion }}'
@@ -65,7 +65,7 @@ jobs:
       display_name: 'feature-gates-regression-bundle VI. - connect + mirrormaker'
       profile: 'azp_connect_mirrormaker'
       cluster_operator_install_type: 'bundle'
-      strimzi_feature_gates: '+ContinueReconciliationOnManualRollingUpdateFailure'
+      strimzi_feature_gates: '-ContinueReconciliationOnManualRollingUpdateFailure'
       strimzi_use_node_pools_in_tests: "false"
       timeout: 360
       releaseVersion: '${{ parameters.releaseVersion }}'
@@ -77,7 +77,7 @@ jobs:
       display_name: 'feature-gates-regression-bundle VII. - remaining system tests'
       profile: 'azp_remaining'
       cluster_operator_install_type: 'bundle'
-      strimzi_feature_gates: '+ContinueReconciliationOnManualRollingUpdateFailure'
+      strimzi_feature_gates: '-ContinueReconciliationOnManualRollingUpdateFailure'
       strimzi_use_node_pools_in_tests: "false"
       timeout: 360
       releaseVersion: '${{ parameters.releaseVersion }}'

diff --git a/.azure/templates/jobs/system-tests/feature_gates_regression_namespace_rbac_jobs.yaml b/.azure/templates/jobs/system-tests/feature_gates_regression_namespace_rbac_jobs.yaml
@@ -8,7 +8,7 @@ jobs:
       cluster_operator_install_type: 'bundle'
       timeout: 360
       strimzi_rbac_scope: NAMESPACE
-      strimzi_feature_gates: '+ContinueReconciliationOnManualRollingUpdateFailure'
+      strimzi_feature_gates: '-ContinueReconciliationOnManualRollingUpdateFailure'
       strimzi_use_node_pools_in_tests: "false"
       releaseVersion: '${{ parameters.releaseVersion }}'
       kafkaVersion: '${{ parameters.kafkaVersion }}'
@@ -22,7 +22,7 @@ jobs:
       cluster_operator_install_type: 'bundle'
       timeout: 360
       strimzi_rbac_scope: NAMESPACE
-      strimzi_feature_gates: '+ContinueReconciliationOnManualRollingUpdateFailure'
+      strimzi_feature_gates: '-ContinueReconciliationOnManualRollingUpdateFailure'
       strimzi_use_node_pools_in_tests: "false"
       releaseVersion: '${{ parameters.releaseVersion }}'
       kafkaVersion: '${{ parameters.kafkaVersion }}'
@@ -36,7 +36,7 @@ jobs:
       cluster_operator_install_type: 'bundle'
       timeout: 360
       strimzi_rbac_scope: NAMESPACE
-      strimzi_feature_gates: '+ContinueReconciliationOnManualRollingUpdateFailure'
+      strimzi_feature_gates: '-ContinueReconciliationOnManualRollingUpdateFailure'
       strimzi_use_node_pools_in_tests: "false"
       releaseVersion: '${{ parameters.releaseVersion }}'
       kafkaVersion: '${{ parameters.kafkaVersion }}'
@@ -50,7 +50,7 @@ jobs:
       cluster_operator_install_type: 'bundle'
       timeout: 360
       strimzi_rbac_scope: NAMESPACE
-      strimzi_feature_gates: '+ContinueReconciliationOnManualRollingUpdateFailure'
+      strimzi_feature_gates: '-ContinueReconciliationOnManualRollingUpdateFailure'
       strimzi_use_node_pools_in_tests: "false"
       releaseVersion: '${{ parameters.releaseVersion }}'
       kafkaVersion: '${{ parameters.kafkaVersion }}'
@@ -64,7 +64,7 @@ jobs:
       cluster_operator_install_type: 'bundle'
       timeout: 360
       strimzi_rbac_scope: NAMESPACE
-      strimzi_feature_gates: '+ContinueReconciliationOnManualRollingUpdateFailure'
+      strimzi_feature_gates: '-ContinueReconciliationOnManualRollingUpdateFailure'
       strimzi_use_node_pools_in_tests: "false"
       releaseVersion: '${{ parameters.releaseVersion }}'
       kafkaVersion: '${{ parameters.kafkaVersion }}'
@@ -78,7 +78,7 @@ jobs:
       cluster_operator_install_type: 'bundle'
       timeout: 360
       strimzi_rbac_scope: NAMESPACE
-      strimzi_feature_gates: '+ContinueReconciliationOnManualRollingUpdateFailure'
+      strimzi_feature_gates: '-ContinueReconciliationOnManualRollingUpdateFailure'
       strimzi_use_node_pools_in_tests: "false"
       releaseVersion: '${{ parameters.releaseVersion }}'
       kafkaVersion: '${{ parameters.kafkaVersion }}'
@@ -92,7 +92,7 @@ jobs:
       cluster_operator_install_type: 'bundle'
       timeout: 360
       strimzi_rbac_scope: NAMESPACE
-      strimzi_feature_gates: '+ContinueReconciliationOnManualRollingUpdateFailure'
+      strimzi_feature_gates: '-ContinueReconciliationOnManualRollingUpdateFailure'
       strimzi_use_node_pools_in_tests: "false"
       releaseVersion: '${{ parameters.releaseVersion }}'
       kafkaVersion: '${{ parameters.kafkaVersion }}'
diff --git a/CHANGELOG.md b/CHANGELOG.md
@@ -3,6 +3,8 @@
 ## 0.44.0
 
 * Add the "Unmanaged" KafkaTopic status update.
+* The `ContinueReconciliationOnManualRollingUpdateFailure` feature gate moves to beta stage and is enabled by default.
+  If needed, `ContinueReconciliationOnManualRollingUpdateFailure` can be disabled in the feature gates configuration in the Cluster Operator.
 
 ### Changes, deprecations and removals
 

diff --git a/cluster-operator/src/test/java/io/strimzi/operator/cluster/ClusterOperatorConfigTest.java b/cluster-operator/src/test/java/io/strimzi/operator/cluster/ClusterOperatorConfigTest.java
@@ -44,7 +44,7 @@ public class ClusterOperatorConfigTest {
         ENV_VARS.put(ClusterOperatorConfig.STRIMZI_KAFKA_MIRROR_MAKER_IMAGES, KafkaVersionTestUtils.getKafkaMirrorMakerImagesEnvVarString());
         ENV_VARS.put(ClusterOperatorConfig.STRIMZI_KAFKA_MIRROR_MAKER_2_IMAGES, KafkaVersionTestUtils.getKafkaMirrorMaker2ImagesEnvVarString());
         ENV_VARS.put(ClusterOperatorConfig.OPERATOR_NAMESPACE.key(), "operator-namespace");
-        ENV_VARS.put(ClusterOperatorConfig.FEATURE_GATES.key(), "+ContinueReconciliationOnManualRollingUpdateFailure");
+        ENV_VARS.put(ClusterOperatorConfig.FEATURE_GATES.key(), "-ContinueReconciliationOnManualRollingUpdateFailure");
         ENV_VARS.put(ClusterOperatorConfig.DNS_CACHE_TTL.key(), "10");
         ENV_VARS.put(ClusterOperatorConfig.POD_SECURITY_PROVIDER_CLASS.key(), "my.package.CustomPodSecurityProvider");
     }
@@ -100,7 +100,7 @@ public void testEnvVars() {
         assertThat(config.getOperationTimeoutMs(), is(30_000L));
         assertThat(config.getConnectBuildTimeoutMs(), is(40_000L));
         assertThat(config.getOperatorNamespace(), is("operator-namespace"));
-        assertThat(config.featureGates().continueOnManualRUFailureEnabled(), is(true));
+        assertThat(config.featureGates().continueOnManualRUFailureEnabled(), is(false));
         assertThat(config.getDnsCacheTtlSec(), is(10));
         assertThat(config.getPodSecurityProviderClass(), is("my.package.CustomPodSecurityProvider"));
     }
@@ -117,6 +117,7 @@ public void testEnvVarsDefault() {
         assertThat(config.getOperationTimeoutMs(), is(Long.parseLong(ClusterOperatorConfig.OPERATION_TIMEOUT_MS.defaultValue())));
         assertThat(config.getOperatorNamespace(), is(nullValue()));
         assertThat(config.getOperatorNamespaceLabels(), is(nullValue()));
+        assertThat(config.featureGates().continueOnManualRUFailureEnabled(), is(true));
         assertThat(config.getDnsCacheTtlSec(), is(Integer.parseInt(ClusterOperatorConfig.DNS_CACHE_TTL.defaultValue())));
         assertThat(config.getPodSecurityProviderClass(), is(ClusterOperatorConfig.POD_SECURITY_PROVIDER_CLASS.defaultValue()));
     }

diff --git a/cluster-operator/src/test/java/io/strimzi/operator/cluster/model/EntityOperatorTest.java b/cluster-operator/src/test/java/io/strimzi/operator/cluster/model/EntityOperatorTest.java
@@ -785,14 +785,14 @@ public void testUserOperatorAndTopicOperatorNetworkPolicy() {
     @ParallelTest
     public void testFeatureGateEnvVars() {
         ClusterOperatorConfig config = new ClusterOperatorConfig.ClusterOperatorConfigBuilder(ResourceUtils.dummyClusterOperatorConfig(), VERSIONS)
-                .with(ClusterOperatorConfig.FEATURE_GATES.key(), "+ContinueReconciliationOnManualRollingUpdateFailure")
+                .with(ClusterOperatorConfig.FEATURE_GATES.key(), "-ContinueReconciliationOnManualRollingUpdateFailure")
                 .build();
 
         EntityOperator eo = EntityOperator.fromCrd(new Reconciliation("test", KAFKA.getKind(), KAFKA.getMetadata().getNamespace(), KAFKA.getMetadata().getName()), KAFKA, SHARED_ENV_PROVIDER, config);
         Deployment dep = eo.generateDeployment(Map.of(), false, null, null);
 
-        assertThat(dep.getSpec().getTemplate().getSpec().getContainers().get(0).getEnv().stream().filter(env -> "STRIMZI_FEATURE_GATES".equals(env.getName())).map(EnvVar::getValue).findFirst().orElseThrow(), is("+ContinueReconciliationOnManualRollingUpdateFailure"));
-        assertThat(dep.getSpec().getTemplate().getSpec().getContainers().get(1).getEnv().stream().filter(env -> "STRIMZI_FEATURE_GATES".equals(env.getName())).map(EnvVar::getValue).findFirst().orElseThrow(), is("+ContinueReconciliationOnManualRollingUpdateFailure"));
+        assertThat(dep.getSpec().getTemplate().getSpec().getContainers().get(0).getEnv().stream().filter(env -> "STRIMZI_FEATURE_GATES".equals(env.getName())).map(EnvVar::getValue).findFirst().orElseThrow(), is("-ContinueReconciliationOnManualRollingUpdateFailure"));
+        assertThat(dep.getSpec().getTemplate().getSpec().getContainers().get(1).getEnv().stream().filter(env -> "STRIMZI_FEATURE_GATES".equals(env.getName())).map(EnvVar::getValue).findFirst().orElseThrow(), is("-ContinueReconciliationOnManualRollingUpdateFailure"));
     }
 
     ////////////////////

diff --git a/...mzi/operator/cluster/operator/assembly/KafkaAssemblyOperatorManualRollingUpdatesTest.java b/...mzi/operator/cluster/operator/assembly/KafkaAssemblyOperatorManualRollingUpdatesTest.java
@@ -357,12 +357,12 @@ public void testManualPodRollingUpdateWithPodSets(VertxTestContext context) {
 
     @Test
     public void testManualPodRollingUpdateWithPodSetsWithError1(VertxTestContext context) {
-        testManualPodRollingUpdateWithPodSetsWithErrorConditions(context, "", true);
+        testManualPodRollingUpdateWithPodSetsWithErrorConditions(context, "-ContinueReconciliationOnManualRollingUpdateFailure", true);
     }
 
     @Test
     public void testManualPodRollingUpdateWithPodSetsWithError3(VertxTestContext context) {
-        testManualPodRollingUpdateWithPodSetsWithErrorConditions(context, "+ContinueReconciliationOnManualRollingUpdateFailure", false);
+        testManualPodRollingUpdateWithPodSetsWithErrorConditions(context, "", false);
     }
 
     private void testManualPodRollingUpdateWithPodSetsWithErrorConditions(VertxTestContext context, String featureGates, boolean expectError) {

diff --git a/...ator/cluster/operator/assembly/KafkaAssemblyOperatorManualRollingUpdatesZooBasedTest.java b/...ator/cluster/operator/assembly/KafkaAssemblyOperatorManualRollingUpdatesZooBasedTest.java
@@ -433,25 +433,25 @@ vertx, new PlatformFeaturesAvailability(false, KUBERNETES_VERSION),
     @Test
     public void testManualPodRollingUpdateWithPodSetsWithError1(VertxTestContext context) {
         testManualPodRollingUpdateWithPodSetsWithErrorConditions(
-            context, false, true, "", true);
+            context, false, true, "-ContinueReconciliationOnManualRollingUpdateFailure", true);
     }
 
     @Test
     public void testManualPodRollingUpdateWithPodSetsWithError2(VertxTestContext context) {
         testManualPodRollingUpdateWithPodSetsWithErrorConditions(
-            context, true, false, "", true);
+            context, true, false, "-ContinueReconciliationOnManualRollingUpdateFailure", true);
     }
 
     @Test
     public void testManualPodRollingUpdateWithPodSetsWithError3(VertxTestContext context) {
         testManualPodRollingUpdateWithPodSetsWithErrorConditions(
-            context, false, true, "+ContinueReconciliationOnManualRollingUpdateFailure", false);
+            context, false, true, "", false);
     }
 
     @Test
     public void testManualPodRollingUpdateWithPodSetsWithError4(VertxTestContext context) {
         testManualPodRollingUpdateWithPodSetsWithErrorConditions(
-            context, true, false, "+ContinueReconciliationOnManualRollingUpdateFailure", false);
+            context, true, false, "", false);
     }
 
     private void testManualPodRollingUpdateWithPodSetsWithErrorConditions(VertxTestContext context,

diff --git a/...io/strimzi/operator/cluster/operator/assembly/KafkaConnectAssemblyOperatorPodSetTest.java b/...io/strimzi/operator/cluster/operator/assembly/KafkaConnectAssemblyOperatorPodSetTest.java
@@ -2230,11 +2230,12 @@ public void testFailingManualRollingUpdate(VertxTestContext context)  {
         ArgumentCaptor<KafkaConnect> connectCaptor = ArgumentCaptor.forClass(KafkaConnect.class);
         when(mockConnectOps.updateStatusAsync(any(), connectCaptor.capture())).thenReturn(Future.succeededFuture());
 
+        ClusterOperatorConfig coConfig = new ClusterOperatorConfig.ClusterOperatorConfigBuilder(ResourceUtils.dummyClusterOperatorConfig(), VERSIONS).with(ClusterOperatorConfig.FEATURE_GATES.key(), "-ContinueReconciliationOnManualRollingUpdateFailure").build();
         KafkaConnectAssemblyOperator ops = new KafkaConnectAssemblyOperator(
                 vertx,
                 new PlatformFeaturesAvailability(false, KUBERNETES_VERSION),
                 supplier,
-                ResourceUtils.dummyClusterOperatorConfig()
+                coConfig
         );
 
         Checkpoint async = context.checkpoint();
@@ -2320,12 +2321,11 @@ public void testManualRollingUpdateWithSuppressedFailure(VertxTestContext contex
         ArgumentCaptor<KafkaConnect> connectCaptor = ArgumentCaptor.forClass(KafkaConnect.class);
         when(mockConnectOps.updateStatusAsync(any(), connectCaptor.capture())).thenReturn(Future.succeededFuture());
 
-        ClusterOperatorConfig coConfig = new ClusterOperatorConfig.ClusterOperatorConfigBuilder(ResourceUtils.dummyClusterOperatorConfig(), VERSIONS).with(ClusterOperatorConfig.FEATURE_GATES.key(), "+ContinueReconciliationOnManualRollingUpdateFailure").build();
         KafkaConnectAssemblyOperator ops = new KafkaConnectAssemblyOperator(
                 vertx,
                 new PlatformFeaturesAvailability(false, KUBERNETES_VERSION),
                 supplier,
-                coConfig
+                ResourceUtils.dummyClusterOperatorConfig()
         );
 
         Checkpoint async = context.checkpoint();

diff --git a/...rimzi/operator/cluster/operator/assembly/KafkaMirrorMaker2AssemblyOperatorPodSetTest.java b/...rimzi/operator/cluster/operator/assembly/KafkaMirrorMaker2AssemblyOperatorPodSetTest.java
@@ -1523,11 +1523,12 @@ public void testFailingManualRollingUpdate(VertxTestContext context)  {
         when(mockConnectClient.list(any(), anyString(), anyInt())).thenReturn(Future.succeededFuture(emptyList()));
         when(mockConnectClient.updateConnectLoggers(any(), anyString(), anyInt(), anyString(), any(OrderedProperties.class))).thenReturn(Future.succeededFuture());
 
+        ClusterOperatorConfig coConfig = new ClusterOperatorConfig.ClusterOperatorConfigBuilder(ResourceUtils.dummyClusterOperatorConfig(), VERSIONS).with(ClusterOperatorConfig.FEATURE_GATES.key(), "-ContinueReconciliationOnManualRollingUpdateFailure").build();
         KafkaMirrorMaker2AssemblyOperator ops = new KafkaMirrorMaker2AssemblyOperator(
                 vertx,
                 new PlatformFeaturesAvailability(false, KUBERNETES_VERSION),
                 supplier,
-                ResourceUtils.dummyClusterOperatorConfig(),
+                coConfig,
                 x -> mockConnectClient
         );
 
@@ -1619,12 +1620,11 @@ public void testManualRollingUpdateWithSuppressedFailure(VertxTestContext contex
         when(mockConnectClient.list(any(), anyString(), anyInt())).thenReturn(Future.succeededFuture(emptyList()));
         when(mockConnectClient.updateConnectLoggers(any(), anyString(), anyInt(), anyString(), any(OrderedProperties.class))).thenReturn(Future.succeededFuture());
 
-        ClusterOperatorConfig coConfig = new ClusterOperatorConfig.ClusterOperatorConfigBuilder(ResourceUtils.dummyClusterOperatorConfig(), VERSIONS).with(ClusterOperatorConfig.FEATURE_GATES.key(), "+ContinueReconciliationOnManualRollingUpdateFailure").build();
         KafkaMirrorMaker2AssemblyOperator ops = new KafkaMirrorMaker2AssemblyOperator(
                 vertx,
                 new PlatformFeaturesAvailability(false, KUBERNETES_VERSION),
                 supplier,
-                coConfig,
+                ResourceUtils.dummyClusterOperatorConfig(),
                 x -> mockConnectClient
         );
 

diff --git a/documentation/modules/managing/proc-manual-rolling-update-pods.adoc b/documentation/modules/managing/proc-manual-rolling-update-pods.adoc
@@ -71,6 +71,6 @@ kubectl annotate pod <cluster_name>-mirrormaker2-<index_number> strimzi.io/manua
 A rolling update of the annotated `Pod` is triggered, as long as the annotation was detected by the reconciliation process.
 When the rolling update of a pod is complete, the annotation is automatically removed from the `Pod`.
 
-NOTE: If the `ContinueReconciliationOnManualRollingUpdateFailure` feature gate is enabled, reconciliation continues even if the manual rolling update of the cluster fails.
+NOTE: Unless the `ContinueReconciliationOnManualRollingUpdateFailure` feature gate is disabled, reconciliation continues even if the manual rolling update of the cluster fails.
 This allows the Cluster Operator to recover from certain rectifiable situations that can be addressed later in the reconciliation. 
 For example, it can recreate a missing Persistent Volume Claim (PVC) or Persistent Volume (PV) that caused the update to fail.
diff --git a/documentation/modules/operators/ref-operator-cluster-feature-gate-releases.adoc b/documentation/modules/operators/ref-operator-cluster-feature-gate-releases.adoc
@@ -30,7 +30,8 @@ Alpha and beta stage features are removed if they do not prove to be useful.
 * The `UseKRaft` feature gate moved to GA stage in Strimzi 0.42.
   It is now permanently enabled and cannot be disabled.
   To use KRaft (ZooKeeper-less Apache Kafka), you still need to use the `strimzi.io/kraft: enabled` annotation on the `Kafka` custom resources or migrate from an existing ZooKeeper-based cluster.
-* The `ContinueReconciliationOnManualRollingUpdateFailure` feature was introduced in Strimzi 0.41 and is disabled by default.
+* The `ContinueReconciliationOnManualRollingUpdateFailure` feature was introduced in Strimzi 0.41 and moved to beta stage in Strimzi 0.44.0.
+  It is now enabled by default, but can be disabled if needed.
 
 NOTE: Feature gates might be removed when they reach GA. This means that the feature was incorporated into the Strimzi core features and can no longer be disabled.
 
@@ -80,8 +81,8 @@ NOTE: Feature gates might be removed when they reach GA. This means that the fea
 
 ¦`ContinueReconciliationOnManualRollingUpdateFailure`
 ¦0.41
-¦0.43 (planned)
-¦n/a
+¦0.44
+¦0.47 (planned)
 
 |===