You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Is your feature request related to a problem? Please describe.
Occasionally, there could be a bug in the Pinniped supervisor or concierge servers which is triggered by the integration test suite and causes the affected pod to crash and be restarted by Kubernetes. This can lead to situations where we don't notice a bug because Kubernetes very helpfully restarts the container and our tests tolerate some eventual consistency (e.g., using require.Eventually() assertions).
Describe the solution you'd like
We should have some test assertion helper that checks on the Pinniped pods before and after each test to validate that they have not been restarted. This is similar to the library.DumpLogs() helper we have already, and should be a straightforward extension of that code.
Describe alternatives you've considered
We could deploy our code under test in pods with restartPolicy: Never, but we'd need to significantly refactor the deployment YAML.
Are you considering submitting a PR for this feature?
Not immediately, please leave a comment if you'd like to take a shot at this.
How will this project improvement be tested?
It's test code, so we can probably just do some manual testing (e.g., insert a panic() into the supervisor code and make sure the tests fail).
How does this change the current architecture?
This is a minor change.
How will this change be backwards compatible?
This is a minor change to tests, no backwards compatibility concerns.
How will this feature be documented?
No documentation needed.
The text was updated successfully, but these errors were encountered:
Is your feature request related to a problem? Please describe.
Occasionally, there could be a bug in the Pinniped supervisor or concierge servers which is triggered by the integration test suite and causes the affected pod to crash and be restarted by Kubernetes. This can lead to situations where we don't notice a bug because Kubernetes very helpfully restarts the container and our tests tolerate some eventual consistency (e.g., using
require.Eventually()
assertions).Describe the solution you'd like
We should have some test assertion helper that checks on the Pinniped pods before and after each test to validate that they have not been restarted. This is similar to the
library.DumpLogs()
helper we have already, and should be a straightforward extension of that code.Describe alternatives you've considered
We could deploy our code under test in pods with
restartPolicy: Never
, but we'd need to significantly refactor the deployment YAML.Are you considering submitting a PR for this feature?
Not immediately, please leave a comment if you'd like to take a shot at this.
How will this project improvement be tested?
It's test code, so we can probably just do some manual testing (e.g., insert a
panic()
into the supervisor code and make sure the tests fail).How does this change the current architecture?
This is a minor change.
How will this change be backwards compatible?
This is a minor change to tests, no backwards compatibility concerns.
How will this feature be documented?
No documentation needed.
The text was updated successfully, but these errors were encountered: