mnojek commented on a change in pull request #18494:
URL: https://github.com/apache/airflow/pull/18494#discussion_r716376092
##########
File path: airflow/providers/google/cloud/example_dags/example_cloud_sql.py
##########
@@ -48,8 +49,8 @@
from airflow.utils.dates import days_ago
GCP_PROJECT_ID = os.environ.get('GCP_PROJECT_ID', 'example-project')
-INSTANCE_NAME = os.environ.get('GCSQL_MYSQL_INSTANCE_NAME', 'test-mysql')
-INSTANCE_NAME2 = os.environ.get('GCSQL_MYSQL_INSTANCE_NAME2', 'test-mysql2')
+INSTANCE_NAME = os.environ.get('GCSQL_MYSQL_INSTANCE_NAME', 'test-mysql') +
str(random.getrandbits(16))
Review comment:
I added this because the instance name is unique only for the specific
Airflow container so each time I run the test within 1 Airflow run, it's the
same. The problem was that if this system test failed before deleting the
instance, it couldn't be run again because it wanted to create instance with
the same instance while it was still present. That's why I added it. So it
works only when it passes within the first execution.
That is a simple workaround to make each test execution independent but
overall it's not the desired solution. Can you point me where else this value
is used?
Ideally, I would want to have independent tests and also a mechanism that
will cleanup all the values created by the test if it fails in the middle. This
is not the scope of this PR, but in near future I plan to work on that.
##########
File path: airflow/providers/google/cloud/example_dags/example_cloud_sql.py
##########
@@ -48,8 +49,8 @@
from airflow.utils.dates import days_ago
GCP_PROJECT_ID = os.environ.get('GCP_PROJECT_ID', 'example-project')
-INSTANCE_NAME = os.environ.get('GCSQL_MYSQL_INSTANCE_NAME', 'test-mysql')
-INSTANCE_NAME2 = os.environ.get('GCSQL_MYSQL_INSTANCE_NAME2', 'test-mysql2')
+INSTANCE_NAME = os.environ.get('GCSQL_MYSQL_INSTANCE_NAME', 'test-mysql') +
str(random.getrandbits(16))
Review comment:
Thanks for the suggestions, @potiuk @mik-laj! I will not implement them
now but I will definitely use your comments when designing the new approach for
running system tests. Hopefully we can make it better and more reliable 😃
In the meantime I will remove this change and hopefully this will not be the
case in the new approach.
--
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.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
