Fixing test_performance_profile_validation

ocs_ci/helpers/helpers.py

@@ -5366,3 +5366,29 @@ def update_volsync_channel():
  logger.error(
  f"Pod volsync-controller-manager not in {constants.STATUS_RUNNING} after 300 seconds"
  )
+
+
+def verify_performance_profile_change(perf_profile):
+ """
+ Verify that newly applied performance profile got updated in storage cluster
+
+ Args:
+ perf_profile (str): Applied performance profile
+
+ Returns:
+ bool: True in case performance profile is updated, False otherwise
+ """
+ from ocs_ci.ocs.resources.storage_cluster import StorageCluster
+
+ # Importing storage cluster object here to avoid circular dependency
+
+ storage_cluster = StorageCluster(
+ resource_name=config.ENV_DATA["storage_cluster_name"],
+ namespace=config.ENV_DATA["cluster_namespace"],
+ )
+
+ assert (
+ perf_profile == storage_cluster.data["spec"]["resourceProfile"]
+ ), f"Performance profile is not updated successfully to {perf_profile}"
+ logger.info(f"Performance profile successfully got updated to {perf_profile} mode")
+ return True

ocs_ci/ocs/constants.py

@@ -1406,15 +1406,15 @@
 }
 
 PERFORMANCE_PROFILE_REQUEST_CPU_VALUES = {
- "mgr": "1500m'",
- "mon": "1500",
+ "mgr": "1500m",
+ "mon": "1500m",
  "osd": "4",
  "mds": "3",
  "rgw": "2",
 }
 PERFORMANCE_PROFILE_REQUEST_MEMORY_VALUES = {
  "mgr": "2Gi",
- "mon": "2Gi",
+ "mon": "3Gi",
  "osd": "8Gi",
  "mds": "8Gi",
  "rgw": "4Gi",
@@ -1461,7 +1461,7 @@
 }
 PERFORMANCE_PROFILE_MEMORY_LIMIT_VALUES = {
  "mgr": "4Gi",
- "mon": "2Gi",
+ "mon": "3Gi",
  "osd": "8Gi",
  "mds": "8Gi",
  "rgw": "4Gi",

ocs_ci/ocs/resources/storage_cluster.py

@@ -1006,7 +1006,7 @@ def verify_storage_cluster():
  elif storage_cluster.data["spec"].get("resourceProfile") != storage_cluster.data[
  "status"
  ].get("lastAppliedResourceProfile"):
- timeout = 1200
+ timeout = 1800
  else:
  timeout = 600
  storage_cluster.wait_for_phase(phase="Ready", timeout=timeout)

tests/functional/z_cluster/test_performance_profile_validation.py

@@ -4,6 +4,7 @@
 from ocs_ci.framework.pytest_customization.marks import (
  tier4a,
  skipif_ocs_version,
+ ignore_leftovers,
  brown_squad,
 )
 from ocs_ci.framework.testlib import (
@@ -14,12 +15,13 @@
 from ocs_ci.framework import config
 from ocs_ci.ocs.ocp import OCP
 from ocs_ci.ocs.resources.storage_cluster import verify_storage_cluster
+from ocs_ci.helpers.helpers import verify_performance_profile_change
 from ocs_ci.ocs import constants
 from ocs_ci.ocs.resources.pod import (
  get_pods_having_label,
  Pod,
 )
-from ocs_ci.utility.utils import run_cmd
+from ocs_ci.utility.utils import run_cmd, TimeoutSampler
 from ocs_ci.ocs.resources.storage_cluster import StorageCluster
 
 
@@ -36,28 +38,73 @@
 @pytest.mark.polarion_id("OCS-5656")
 @pytest.mark.polarion_id("OCS-5657")
 class TestProfileDefaultValuesCheck(ManageTest):
- def test_validate_cluster_resource_profile(self):
+ @pytest.mark.parametrize(
+ argnames=["perf_profile"],
+ argvalues=[
+ pytest.param(*["performance"]),
+ pytest.param(*["lean"]),
+ pytest.param(*["balanced"]),
+ ],
+ )
+ @ignore_leftovers
+ def test_validate_cluster_resource_profile(self, perf_profile):
  """
  Testcase to validate osd, mgr, mon, mds and rgw pod memory and cpu values
  are matching with the predefined set of values post profile updation
 
  """
  pv_pod_obj = []
+ namespace = config.ENV_DATA["cluster_namespace"]
  log.info("Obtaining the performance profile values from the cluster")
  storage_cluster_name = config.ENV_DATA["storage_cluster_name"]
  storage_cluster = StorageCluster(
  resource_name=storage_cluster_name,
  namespace=config.ENV_DATA["cluster_namespace"],
  )
- performance_profile = storage_cluster.data["spec"]["resourceProfile"]
- if performance_profile == constants.PERFORMANCE_PROFILE_LEAN:
+ self.perf_profile = perf_profile
+ try:
+ exist_performance_profile = storage_cluster.data["spec"]["resourceProfile"]
+ curr_prof = storage_cluster.data["spec"]["resourceProfile"]
+ log.info(f"Current performance profile is {curr_prof}")
+ except KeyError:
+ # On some occasions, a cluster will be deployed without performance profile, In that case, set it to None.
+ log.info(
+ "If a cluster is deployed without performance profile, set existing_profile value as None"
+ )
+ exist_performance_profile = None
+ pass
+ if exist_performance_profile == self.perf_profile:
+ log.info("Performance profile is same as profile that is already present")
+ else:
+ ptch = f'{{"spec": {{"resourceProfile":"{self.perf_profile}"}}}}'
+ ptch_cmd = (
+ f"oc patch storagecluster {storage_cluster.data.get('metadata').get('name')} "
+ f"-n {namespace} --type merge --patch '{ptch}'"
+ )
+ run_cmd(ptch_cmd)
+ log.info("Verify storage cluster is in Ready state")
+ verify_storage_cluster()
+
+ # Wait up to 600 seconds for performance changes to reflect
+ sample = TimeoutSampler(
+ timeout=600,
+ sleep=300,
+ func=verify_performance_profile_change,
+ perf_profile=self.perf_profile,
+ )
+ if not sample.wait_for_func_status(True):
+ raise Exception(
+ f"Performance profile is not updated successfully to {self.perf_profile}"
+ )
+
+ if self.perf_profile == constants.PERFORMANCE_PROFILE_LEAN:
  expected_cpu_request_values = constants.LEAN_PROFILE_REQUEST_CPU_VALUES
  expected_memory_request_values = (
  constants.LEAN_PROFILE_REQUEST_MEMORY_VALUES
  )
  expected_cpu_limit_values = constants.LEAN_PROFILE_CPU_LIMIT_VALUES
  expected_memory_limit_values = constants.LEAN_PROFILE_MEMORY_LIMIT_VALUES
- elif performance_profile == constants.PERFORMANCE_PROFILE_BALANCED:
+ elif self.perf_profile == constants.PERFORMANCE_PROFILE_BALANCED:
  expected_cpu_request_values = constants.BALANCED_PROFILE_REQUEST_CPU_VALUES
  expected_memory_request_values = (
  constants.BALANCED_PROFILE_REQUEST_MEMORY_VALUES
@@ -66,12 +113,12 @@ def test_validate_cluster_resource_profile(self):
  expected_memory_limit_values = (
  constants.BALANCED_PROFILE_MEMORY_LIMIT_VALUES
  )
- elif performance_profile == constants.PERFORMANCE_PROFILE_PERFORMANCE:
+ elif self.perf_profile == constants.PERFORMANCE_PROFILE_PERFORMANCE:
  expected_cpu_request_values = (
  constants.PERFORMANCE_PROFILE_REQUEST_CPU_VALUES
  )
  expected_memory_request_values = (
- constants.PERFORMANCE_PROFILE_REQUESt_LIMIT_VALUES
+ constants.PERFORMANCE_PROFILE_REQUEST_MEMORY_VALUES
  )
  expected_cpu_limit_values = constants.PERFORMANCE_PROFILE_CPU_LIMIT_VALUES
  expected_memory_limit_values = (
@@ -80,8 +127,6 @@ def test_validate_cluster_resource_profile(self):
  else:
  log.error("Does not match any performance profiles")
 
- log.info(performance_profile)
-
  label_selector = list(expected_cpu_limit_values.keys())
 
  for label in label_selector:
@@ -90,11 +135,17 @@ def test_validate_cluster_resource_profile(self):
  ):
  pv_pod_obj.append(Pod(**pod))
  podd = Pod(**pod)
- log.info(podd.name)
+ log.info(f"Verifying memory and cpu values for pod {podd.name}")
+ log.info(f"RequestCPU{expected_cpu_request_values}")
+ log.info(f"LimitCPU{expected_cpu_limit_values}")
+ log.info(f"RequestMEM{expected_memory_request_values}")
+ log.info(f"LimitMEM{expected_memory_limit_values}")
  resource_dict = OCP(
  namespace=config.ENV_DATA["cluster_namespace"], kind="pod"
  ).get(resource_name=podd.name)["spec"]["containers"][0]["resources"]
-
+ log.info(
+ f"CPU request and limit values for pod {podd.name} are {resource_dict}"
+ )
  assert (
  resource_dict["limits"]["cpu"] == expected_cpu_limit_values[label]
  and resource_dict["limits"]["memory"]
@@ -114,6 +165,7 @@ def test_validate_cluster_resource_profile(self):
  pytest.param(*["balanced"], marks=pytest.mark.polarion_id("OCS-5644")),
  ],
  )
+ @ignore_leftovers
  def test_change_cluster_resource_profile(self, perf_profile):
  """
  Testcase to validate osd, mgr, mon, mds and rgw pod memory and cpu values
@@ -128,7 +180,15 @@ def test_change_cluster_resource_profile(self, perf_profile):
  resource_name=storage_cluster_name,
  namespace=config.ENV_DATA["cluster_namespace"],
  )
- exist_performance_profile = storage_cluster.data["spec"]["resourceProfile"]
+ try:
+ exist_performance_profile = storage_cluster.data["spec"]["resourceProfile"]
+ except KeyError:
+ # On some occasions, a cluster will be deployed without performance profile, In that case, set it to None.
+ log.info(
+ "If a cluster is deployed without performance profile, set existing_profile value as None"
+ )
+ exist_performance_profile = None
+ pass
  if exist_performance_profile == self.perf_profile:
  log.info("Performance profile is same as profile that is already present")
  else:
@@ -142,14 +202,17 @@ def test_change_cluster_resource_profile(self, perf_profile):
 
  verify_storage_cluster()
 
- # Verify that storage cluster is updated with new profile
-
- assert (
- self.perf_profile == storage_cluster.data["spec"]["resourceProfile"]
- ), f"Performance profile is not updated successfully to {self.perf_profile}"
- log.info(
- f"Performance profile successfully got updated to {self.perf_profile} mode"
+ sample = TimeoutSampler(
+ timeout=600,
+ sleep=30,
+ func=verify_performance_profile_change,
+ perf_profile=self.perf_profile,
  )
+ if not sample.wait_for_func_status(True):
+ raise Exception(
+ f"Performance profile is not updated successfully to {self.perf_profile}"
+ )
+
  log.info("Reverting profile changes")
  ptch = f'{{"spec": {{"resourceProfile":"{exist_performance_profile}"}}}}'