change work directory

Signed-off-by: helenxie-bit <[email protected]>

.github/workflows/e2e-test-tune-api.yaml

@@ -16,24 +16,6 @@ jobs:
     steps:
       - name: Checkout
         uses: actions/checkout@v4
-
-      - name: Move docker data directory
-        shell: bash
-        run: |
-          echo "Stopping docker service ..."
-          sudo systemctl stop docker
-          DOCKER_DEFAULT_ROOT_DIR=/var/lib/docker
-          DOCKER_ROOT_DIR=/mnt/docker
-          echo "Moving ${DOCKER_DEFAULT_ROOT_DIR} -> ${DOCKER_ROOT_DIR}"
-          sudo mv ${DOCKER_DEFAULT_ROOT_DIR} ${DOCKER_ROOT_DIR}
-          echo "Creating symlink ${DOCKER_DEFAULT_ROOT_DIR} -> ${DOCKER_ROOT_DIR}"
-          sudo ln -s ${DOCKER_ROOT_DIR} ${DOCKER_DEFAULT_ROOT_DIR}
-          echo "$(sudo ls -l ${DOCKER_DEFAULT_ROOT_DIR})"
-          echo "Starting docker service ..."
-          sudo systemctl daemon-reload
-          sudo systemctl start docker
-          echo "Docker service status:"
-          sudo systemctl --no-pager -l -o short status docker
 
       - name: Setup Test Env
         uses: ./.github/workflows/template-setup-e2e-test

.github/workflows/template-e2e-test/action.yaml

@@ -47,3 +47,4 @@ runs:
         else
           ./test/e2e/v1beta1/scripts/gh-actions/run-e2e-experiment.sh ${{ inputs.experiments }}
         fi
+      working-directory: /mnt/docker

test/e2e/v1beta1/scripts/gh-actions/run-e2e-tune-api.py

@@ -36,6 +36,25 @@ def get_experiment_pods_logs(katib_client: KatibClient, exp_name: str, exp_names
                     namespace=exp_namespace,
                     container="metrics-logger-and-collector"  # Specify the desired container
                 )
+                logging.info(f"Logs for pod {pod.metadata.name}:\n{pod_logs}")
+            except Exception as e:
+                logging.error(f"Failed to get logs for pod {pod.metadata.name}: {str(e)}")
+
+def get_experiment_pods_logs_2(katib_client: KatibClient, exp_name: str, exp_namespace: str):
+    # List all the pods in the namespace
+    v1 = client.CoreV1Api()
+    pods = v1.list_namespaced_pod(namespace=exp_namespace)
+
+    # Filter pods related to the specific Katib Experiment
+    for pod in pods.items:
+        if exp_name in pod.metadata.name:
+            logging.info(f"Fetching logs for pod: {pod.metadata.name}")
+            try:
+                # Specify the container name when retrieving logs
+                pod_logs = v1.read_namespaced_pod_log(
+                    name=pod.metadata.name,
+                    namespace=exp_namespace,
+                )
                 logging.info(f"Logs for pod {pod.metadata.name} (container: metrics-logger-and-collector):\n{pod_logs}")
             except Exception as e:
                 logging.error(f"Failed to get logs for pod {pod.metadata.name}: {str(e)}")
@@ -197,7 +216,7 @@ def run_e2e_experiment_create_by_tune_with_external_model(
         logging.info("---------------------------------------------------------------")
         logging.info(f"E2E is failed for Experiment created by tune: {exp_namespace}/{exp_name}-2")
         get_experiment_pods_logs(katib_client, f"{exp_name}-2", exp_namespace)
-        get_experiment_pods_logs(katib_client, "katib-controller", "kubeflow")
+        get_experiment_pods_logs_2(katib_client, "katib-controller", "kubeflow")
         raise e
     finally:
         # Delete the Experiment.