refactored virt_api_pod, moved wait_for_pod_running_by_prefix to tests/utils.py to reuse it

Author: albarker-rh

tests/install_upgrade_operators/product_install/utils.py

@@ -3,17 +3,14 @@
 import subprocess
 
 from ocp_resources.deployment import Deployment
-from ocp_utilities.operators import TIMEOUT_5MIN
-from timeout_sampler import TimeoutExpiredError, TimeoutSampler
 
+from tests.utils import wait_for_pod_running_by_prefix
 from utilities.constants import (
     HOSTPATH_PROVISIONER,
     HOSTPATH_PROVISIONER_OPERATOR,
-    TIMEOUT_5SEC,
     TIMEOUT_15MIN,
 )
 from utilities.data_collector import write_to_file
-from utilities.infra import get_not_running_pods, get_pod_by_name_prefix
 from utilities.storage import verify_hpp_pool_health
 
 PRINT_COMMAND = '{printf "%s%s",sep,$0;sep=","}'
@@ -71,49 +68,6 @@ def get_all_resources(file_name, base_directory):
     }
 
 
-def wait_for_pod_running_by_prefix(
-    admin_client,
-    namespace_name,
-    pod_prefix,
-    expected_number_of_pods,
-    number_of_consecutive_checks=3,
-    timeout=TIMEOUT_5MIN,
-):
-    samples = TimeoutSampler(
-        wait_timeout=timeout,
-        sleep=TIMEOUT_5SEC,
-        func=get_pod_by_name_prefix,
-        dyn_client=admin_client,
-        pod_prefix=pod_prefix,
-        namespace=namespace_name,
-        get_all=True,
-    )
-    pod_names = None
-    not_running_pods = None
-    try:
-        current_check = 0
-        for sample in samples:
-            if sample:
-                not_running_pods = get_not_running_pods(pods=sample)
-                pod_names = [pod.name for pod in sample]
-                LOGGER.info(f"All {pod_prefix} pods: {pod_names}, not running: {not_running_pods}")
-                if not_running_pods:
-                    current_check = 0
-                else:
-                    if expected_number_of_pods == len(sample):
-                        current_check += 1
-                    else:
-                        current_check = 0
-            if current_check >= number_of_consecutive_checks:
-                return True
-    except TimeoutExpiredError:
-        LOGGER.error(
-            f"timeout waiting for all {pod_prefix} pods in namespace {namespace_name} to reach "
-            f"running state, out of {pod_names}, following pods are in not running state: {not_running_pods}"
-        )
-        raise
-
-
 def validate_hpp_installation(admin_client, cnv_namespace, schedulable_nodes):
     hpp_deployment = Deployment(name=HOSTPATH_PROVISIONER_OPERATOR, namespace=cnv_namespace.name)
     assert hpp_deployment.exists

tests/observability/metrics/conftest.py

@@ -35,7 +35,7 @@
     wait_for_metric_vmi_request_cpu_cores_output,
 )
 from tests.observability.utils import validate_metrics_value
-from tests.utils import create_vms
+from tests.utils import create_vms, wait_for_pod_running_by_prefix
 from utilities import console
 from utilities.constants import (
     IPV4_STR,
@@ -623,78 +623,29 @@ def aaq_resource_hard_limit_and_used(application_aware_resource_quota):
 
 
 @pytest.fixture()
-def virt_api_pod(admin_client, hco_namespace):
-    """
-    Wait for exactly one virt-api pod to be running.
-
-    Polls for virt-api pods and waits until exactly one pod remains.
-    Pods marked for deletion will be fully deleted before this condition is met.
-
-    Args:
-        admin_client: DynamicClient for cluster operations
-        hco_namespace: HCO namespace object
-
-    Returns:
-        Pod: The single running virt-api pod
-
-    Raises:
-        TimeoutExpiredError: If exactly one pod is not found within timeout period
-    """
-    sample = []
-    try:
-        for sample in TimeoutSampler(
-            wait_timeout=TIMEOUT_2MIN,
-            sleep=TIMEOUT_15SEC,
-            func=get_pod_by_name_prefix,
-            dyn_client=admin_client,
-            pod_prefix="virt-api",
-            namespace=hco_namespace.name,
-            get_all=True,
-        ):
-            if sample and len(sample) == 1:
-                return sample[0]
-    except TimeoutExpiredError:
-        pods_info = [
-            (pod.name, pod.instance.status.phase, pod.instance.metadata.get("deletionTimestamp")) for pod in sample
-        ]
-        LOGGER.exception(f"Should only be 1 virt-api pod running: found {pods_info}")
-        raise
+def virt_api_pod_after_scale_to_one(admin_client, hco_namespace):
+    wait_for_pod_running_by_prefix(
+        admin_client=admin_client, namespace_name=hco_namespace.name, pod_prefix="virt-api", expected_number_of_pods=1
+    )
 
 
 @pytest.fixture()
-def virt_api_initial_metric_value(prometheus, virt_api_pod):
-    """
-    Get the initial value of kubevirt_vm_created_by_pod_total metric for a specific virt-api pod.
-
-    Args:
-        prometheus: Prometheus client fixture
-        virt_api_pod: The virt-api pod to query metrics for
-
-    Returns:
-        int: The current value of the metric for the specified pod
-    """
+def virt_api_initial_metric_value(prometheus, virt_api_pod_after_scale_to_one):
     metric_query = (
-        f"{KUBEVIRT_VM_CREATED_BY_POD_TOTAL}{{pod='{virt_api_pod.name}',namespace='{virt_api_pod.namespace}'}}"
+        f"{KUBEVIRT_VM_CREATED_BY_POD_TOTAL}"
+        f"{{pod='{virt_api_pod_after_scale_to_one.name}',"
+        f"namespace='{virt_api_pod_after_scale_to_one.namespace}'}}"
     )
     return int(get_metrics_value(prometheus=prometheus, metrics_name=metric_query))
 
 
 @pytest.fixture()
-def vm_in_virt_api_ns(virt_api_pod):
-    """
-    Creates a VM in the virt-api namespace
-
-    Args:
-        virt_api_pod: The virt-api pod to fetch the namespace from
-
-    Yields:
-        VM: The created VM
-    """
+def vm_in_virt_api_ns(virt_api_pod_after_scale_to_one):
     vm_name = "virt-api-vm"
 
     with VirtualMachineForTests(
         name=vm_name,
-        namespace=virt_api_pod.namespace,
+        namespace=virt_api_pod_after_scale_to_one.namespace,
         body=fedora_vm_body(name=vm_name),
         ssh=False,
     ) as vm:

tests/observability/metrics/test_vms_metrics.py

@@ -529,12 +529,14 @@ def test_kubevirt_vm_created_by_pod_total(
         prometheus,
         disabled_virt_operator,
         scaled_deployment,
-        virt_api_pod,
+        virt_api_pod_after_scale_to_one,
         virt_api_initial_metric_value,
         vm_in_virt_api_ns,
     ):
         metric_query = (
-            f"{KUBEVIRT_VM_CREATED_BY_POD_TOTAL}{{pod='{virt_api_pod.name}',namespace='{virt_api_pod.namespace}'}}"
+            f"{KUBEVIRT_VM_CREATED_BY_POD_TOTAL}"
+            f"{{pod='{virt_api_pod_after_scale_to_one.name}',"
+            f"namespace='{virt_api_pod_after_scale_to_one.namespace}'}}"
         )
         validate_metrics_value(
             prometheus=prometheus, metric_name=metric_query, expected_value=str(virt_api_initial_metric_value + 1)

tests/utils.py

@@ -18,6 +18,7 @@
 from ocp_resources.resource import ResourceEditor
 from ocp_resources.virtual_machine import VirtualMachine
 from ocp_resources.virtual_machine_instance_migration import VirtualMachineInstanceMigration
+from ocp_utilities.operators import TIMEOUT_5MIN
 from pyhelper_utils.shell import run_ssh_commands
 from timeout_sampler import TimeoutExpiredError, TimeoutSampler
 
@@ -39,6 +40,8 @@
     get_artifactory_header,
     get_artifactory_secret,
     get_http_image_url,
+    get_not_running_pods,
+    get_pod_by_name_prefix,
 )
 from utilities.virt import (
     VirtualMachineForTests,
@@ -564,3 +567,46 @@ def create_cirros_vm(
         if wait_running:
             running_vm(vm=vm, wait_for_interfaces=False)
         yield vm
+
+
+def wait_for_pod_running_by_prefix(
+    admin_client,
+    namespace_name,
+    pod_prefix,
+    expected_number_of_pods,
+    number_of_consecutive_checks=3,
+    timeout=TIMEOUT_5MIN,
+):
+    samples = TimeoutSampler(
+        wait_timeout=timeout,
+        sleep=TIMEOUT_5SEC,
+        func=get_pod_by_name_prefix,
+        dyn_client=admin_client,
+        pod_prefix=pod_prefix,
+        namespace=namespace_name,
+        get_all=True,
+    )
+    pod_names = None
+    not_running_pods = None
+    try:
+        current_check = 0
+        for sample in samples:
+            if sample:
+                not_running_pods = get_not_running_pods(pods=sample)
+                pod_names = [pod.name for pod in sample]
+                LOGGER.info(f"All {pod_prefix} pods: {pod_names}, not running: {not_running_pods}")
+                if not_running_pods:
+                    current_check = 0
+                else:
+                    if expected_number_of_pods == len(sample):
+                        current_check += 1
+                    else:
+                        current_check = 0
+            if current_check >= number_of_consecutive_checks:
+                return True
+    except TimeoutExpiredError:
+        LOGGER.error(
+            f"timeout waiting for all {pod_prefix} pods in namespace {namespace_name} to reach "
+            f"running state, out of {pod_names}, following pods are in not running state: {not_running_pods}"
+        )
+        raise