hunyadi-dev commented on a change in pull request #995:
URL: https://github.com/apache/nifi-minifi-cpp/pull/995#discussion_r572865677
##########
File path: docker/test/integration/MiNiFi_integration_test_driver.py
##########
@@ -0,0 +1,208 @@
+from subprocess import Popen, PIPE, STDOUT
+
+import docker
+import logging
+import os
+import shutil
+import threading
+import time
+import uuid
+
+from pydoc import locate
+
+from minifi.core.InputPort import InputPort
+
+from minifi.core.DockerTestCluster import DockerTestCluster
+from minifi.core.SingleNodeDockerCluster import SingleNodeDockerCluster
+from minifi.core.DockerTestDirectoryBindings import DockerTestDirectoryBindings
+
+from minifi.validators.EmptyFilesOutPutValidator import
EmptyFilesOutPutValidator
+from minifi.validators.NoFileOutPutValidator import NoFileOutPutValidator
+from minifi.validators.SingleFileOutputValidator import
SingleFileOutputValidator
+
+class MiNiFi_integration_test():
+ def __init__(self, context):
+ logging.info("MiNiFi_integration_test init")
+ self.test_id = str(uuid.uuid4())
+ self.clusters = {}
+
+ self.connectable_nodes = []
+ # Remote process groups are not connectables
+ self.remote_process_groups = []
+ self.file_system_observer = None
+
+ self.docker_network = None
+
+ self.docker_directory_bindings = DockerTestDirectoryBindings()
+
self.docker_directory_bindings.create_new_data_directories(self.test_id)
+
+ def __del__(self):
+ logging.info("MiNiFi_integration_test cleanup")
+
+ # Clean up network, for some reason only this order of events work for
cleanup
+ if self.docker_network is not None:
+ logging.info('Cleaning up network network: %s',
self.docker_network.name)
+ while len(self.docker_network.containers) != 0:
+ for container in self.docker_network.containers:
+ self.docker_network.disconnect(container, force=True)
+ self.docker_network.reload()
+ self.docker_network.remove()
+
+ container_ids = []
+ for cluster in self.clusters.values():
+ for container in cluster.containers.values():
+ container_ids.append(container.id)
+ del cluster
+
+ # Backup for cleaning up containers as the cluster deleter is not
reliable
+ docker_client = docker.from_env()
+ for container_id in container_ids:
+ wait_start_time = time.perf_counter()
+ while (time.perf_counter() - wait_start_time) < 35:
+ # There is no clean way to check for container existence
+ try:
+ container = docker_client.containers.get(container_id)
+ logging.error("Failure when trying to clean up containers.
Attempting secondary cleanup.")
+ container.kill()
+ time.sleep(5)
+ container.remove(v=True, force=True)
+ time.sleep(5)
+ except docker.errors.NotFound:
+ break
+ try:
+ container = docker_client.containers.get(container_id)
+ logging.error("All attempts to clean up docker containers were
unsuccessful.")
+ except docker.errors.NotFound:
+ logging.info("Docker container secondary cleanup successful.")
+ pass
Review comment:
All right, but the problem is that you do not know if the removal
succeeded unless you query for the same container again. The extracted function
would look like this:
```python
def delete_docker_container_by_id(container_id):
try:
container = docker_client.containers.get(container_id)
container.remove(v=True, force=True)
except docker.errors.NotFound:
logging.error("Attempt to clean up a contaner that has already been
removed before.")
return True
try:
docker_client.containers.get(container_id)
logging.error("Failure when trying to clean up containers %s.",
container_id)
return False
except docker.errors.NotFound:
logging.info("Docker container removal successful for container %d",
container_id)
return True
```
This looks just as bad does it not?
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org
