-
Notifications
You must be signed in to change notification settings - Fork 661
Remove ZooKeeperClient/KafkaZkClient usage from tests #893
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Conversation
These are internal classes and hence can change in incompatible ways. One such incompatible change committed to Apache 3.0 is the reason why we need this in 7.0.x too.
One test failed for build 2:
This looks to be flaky, it fails occasionally without this change. And it passed locally after a few retries. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
LGTM assuming builds pass.
* origin/7.0.x: Remove ZooKeeperClient/KafkaZkClient usage from tests (#893)
This seems to have caused all integration tests based on |
Recent AK changes related to the removal of ZooKeeper have necessitated the accompanying changes in kafka-rest's tests. confluentinc#893 has done that, but unfortunately it also removed the initialization of the zkConnect field which then caused an NPE when it ended up being added to a ConcurrentHashMap as a null. This attempts to fix the problem by the simplest of approaches: putting back the initialization of zkConnect.
Recent AK changes related to the removal of ZooKeeper have necessitated the accompanying changes in kafka-rest's tests. #893 has done that, but unfortunately it also removed the initialization of the zkConnect field which then caused an NPE when it ended up being added to a ConcurrentHashMap as a null. This attempts to fix the problem by the simplest of approaches: putting back the initialization of zkConnect.
The two tests ignored here are way too flaky and we have seen at least one occasion were one of them was hiding an even worse problem: confluentinc#893. We have added tasks for actually fixing them, but in the meantime they should not be causing as much trouble as they currently are.
Sorry for that. Can we fix the build not to ignore integration tests in such cases? |
I don't think you (or anyone) could have done anything better in this specific situation. As for not ignoring other phases, we should very much look into that. |
The two tests ignored here are way too flaky and we have seen at least one occasion were one of them was hiding an even worse problem: #893. We have added tasks for actually fixing them, but in the meantime they should not be causing as much trouble as they currently are.
These are internal classes and hence can change in incompatible ways.
One such incompatible change committed to Apache Kafka 3.0 is the reason
why we need this in 7.0.x too.