Merge branch 'master' into snapshot-delete

e2e/requirements.txt

@@ -4,5 +4,5 @@ directio==1.3
 flake8
 kubernetes==27.2.0
 requests==2.31.0
-boto3==1.34.49
+boto3==1.34.51
 pyyaml==6.0.1

manager/integration/tests/test_basic.py

@@ -101,6 +101,7 @@
 from common import create_and_wait_deployment
 from common import get_custom_object_api_client
 from common import RETRY_COUNTS_SHORT
+from common import scale_up_engine_image_daemonset
 
 from backupstore import backupstore_delete_volume_cfg_file
 from backupstore import backupstore_cleanup
@@ -4426,10 +4427,7 @@ def test_backuptarget_available_during_engine_image_not_ready(client, apps_api):
             common.wait_for_backup_target_available(client, False)
 
             # Scale up the engine image DaemonSet
-            body = [{"op": "remove",
-                     "path": "/spec/template/spec/nodeSelector/foo"}]
-            apps_api.patch_namespaced_daemon_set(
-                name=ds_name, namespace='longhorn-system', body=body)
+            scale_up_engine_image_daemonset(client)
             common.wait_for_backup_target_available(client, True)
 
             # Sleep 1 second to prevent the same time

manager/integration/tests/test_metric.py

@@ -8,6 +8,7 @@
 from common import client, core_api, volume_name  # NOQA
 
 from common import delete_replica_processes
+from common import check_volume_data
 from common import create_pv_for_volume
 from common import create_pvc_for_volume
 from common import create_snapshot
@@ -194,6 +195,21 @@ def filter_metric_by_labels(metrics, labels):
         assert total_metrics["value"] >= 0.0
 
 
+def wait_for_metric_volume_actual_size(core_api, metric_name, metric_labels, actual_size): # NOQA
+    for _ in range(RETRY_COUNTS):
+        time.sleep(RETRY_INTERVAL)
+
+        try:
+            check_metric(core_api, metric_name,
+                         metric_labels, actual_size)
+            return
+        except AssertionError:
+            continue
+
+    check_metric(core_api, metric_name,
+                 metric_labels, actual_size)
+
+
 def wait_for_metric_count_all_nodes(client, core_api, metric_name, metric_labels, expected_count): # NOQA
     for _ in range(RETRY_COUNTS):
         time.sleep(RETRY_INTERVAL)
@@ -271,7 +287,8 @@ def test_volume_metrics(client, core_api, volume_name, pvc_namespace): # NOQA
     volume = client.by_id_volume(volume_name)
     volume.attach(hostId=lht_hostId)
     volume = wait_for_volume_healthy(client, volume_name)
-    write_volume_random_data(volume)
+    data = write_volume_random_data(volume)
+    check_volume_data(volume, data)
     volume = client.by_id_volume(volume_name)
     actual_size = float(volume.controllers[0].actualSize)
     capacity_size = float(volume.size)
@@ -284,8 +301,9 @@ def test_volume_metrics(client, core_api, volume_name, pvc_namespace): # NOQA
     }
 
     # check volume metric basic
-    check_metric(core_api, "longhorn_volume_actual_size_bytes",
-                 metric_labels, actual_size)
+    wait_for_metric_volume_actual_size(core_api,
+                                       "longhorn_volume_actual_size_bytes",
+                                       metric_labels, actual_size)
     check_metric(core_api, "longhorn_volume_capacity_bytes",
                  metric_labels, capacity_size)
     check_metric(core_api, "longhorn_volume_read_throughput",

manager/integration/tests/test_scheduling.py

@@ -463,6 +463,7 @@ def test_replica_rebuild_per_volume_limit(client, core_api, storage_class, sts_n
 
     vol = common.wait_for_volume_replicas_mode(client, vol_name, 'RW',
                                                replica_count=r_count)
+    wait_for_volume_healthy(client, vol_name)
 
     # Delete 4 volume replicas
     del vol.replicas[0]

test_framework/scripts/longhorn-setup.sh

@@ -341,6 +341,9 @@ run_longhorn_upgrade_test(){
 
   # get upgrade test junit xml report
   kubectl cp ${LONGHORN_UPGRADE_TEST_POD_NAME}:${LONGHORN_JUNIT_REPORT_PATH} "${TF_VAR_tf_workspace}/${LONGHORN_UPGRADE_TEST_POD_NAME}-junit-report.xml" -c longhorn-test-report
+
+  # delete upgrade test pod
+  kubectl delete -f ${LONGHORN_UPGRADE_TESTS_MANIFEST_FILE_PATH}
 }