e2e: enable "TestSingleNodeFailure" on k8s

Author: hamdy-simplyblock

.github/workflows/k8s-e2e.yaml

@@ -223,8 +223,8 @@ jobs:
       - name: Install Helm Chart for spdk-csi
         run: |
           cd $GITHUB_WORKSPACE/spdk-csi/charts/latest/spdk-csi/
-          echo "Sleeping for 120 seconds before helm install"
-          sleep 120
+          echo "Sleeping for 30 seconds before helm install"
+          sleep 30
           helm install spdk-csi ./ \
             --namespace spdk-csi \
             --create-namespace \

e2e/__init__.py

@@ -34,6 +34,4 @@ def get_all_tests(custom=True, k8s_test=False):
         tests.remove(TestLvolFioNpcs0)
         tests.remove(TestLvolFioNpcs1)
         tests.remove(TestLvolFioNpcs2)
-    if k8s_test:
-        tests.remove(TestSingleNodeFailure)
     return tests

e2e/e2e_tests/ha_journal/lvol_journal_device_node_restart.py

@@ -262,7 +262,7 @@ def run(self):
         # Step 13: Ungracefully stop node 1 (container shutdown)
         self.logger.info("Stopping container on node 1 (ungraceful shutdown).")
         node_ip = self.journal_manager.sn_journal_map[self.lvol_sn_node]['primary_journal'][1]
-        self.ssh_obj.stop_docker_containers(node_ip, "spdk")
+        self.ssh_obj.stop_spdk_process(node_ip)
         self.sbcli_utils.wait_for_storage_node_status(self.lvol_sn_node,
                                                       "unreachable",
                                                       timeout=300)
@@ -377,7 +377,7 @@ def stop_and_restart_based_on_journals(self):
             # Step 5: Force stop the node with secondary journal 2
             secondary_node_2 = self.extract_node_from_journal(secondary_journal_2[0])
             self.logger.info(f"Forcefully stopping node: {secondary_node_2} (secondary journal 2)")
-            self.ssh_obj.stop_docker_containers(node=secondary_journal_2[1], container_name="spdk")
+            self.ssh_obj.stop_spdk_process(node=secondary_journal_2[1])
             sleep_n_sec(420)
 
             self.sbcli_utils.restart_node(node_uuid=secondary_node_2)

e2e/e2e_tests/single_node_failure.py

@@ -110,18 +110,14 @@ def run(self):
                          )
 
         sleep_n_sec(30)
-        if not self.k8s_test:
-            self.ssh_obj.stop_docker_containers(node=node_ip, container_name="spdk")
-        else:
-            # TODO: Add K8s deployment delete step
-            pass
-        
+        self.ssh_obj.stop_spdk_process(node=node_ip)
+
         try:
             self.logger.info(f"Waiting for node to become offline/unreachable, {no_lvol_node_uuid}")
             self.sbcli_utils.wait_for_storage_node_status(no_lvol_node_uuid,
                                                           ["unreachable", "offline"],
                                                           timeout=500)
-            sleep_n_sec(30)
+            # sleep_n_sec(30)
             # self.validations(node_uuid=no_lvol_node_uuid,
             #                 node_status=["offline", "in_shutdown", "in_restart"],
             #                 # The status changes between them very quickly hence

e2e/utils/ssh_utils.py

@@ -340,18 +340,14 @@ def list_files(self, node, location):
         output, error = self.exec_command(node=node, command=cmd)
         return output
 
-    def stop_docker_containers(self, node, container_name=None):
-        """Stops given docker container. Stops all if no name is given
+    def stop_spdk_process(self, node):
+        """Stops spdk process
 
         Args:
             node (str): Node IP
-            container_name (str): Name of container to stop
         """
-        if container_name:
-            cmd = f"sudo docker stop {container_name}"
-        else:
-            cmd = "sudo docker stop $(sudo docker ps -a -q)"
-        
+
+        cmd = "curl 0.0.0.0:5000/snode/spdk_process_kill"
         output, error = self.exec_command(node=node, command=cmd)
         return output