[
https://issues.apache.org/jira/browse/KAFKA-17174?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17868220#comment-17868220
]
Chia-Ping Tsai commented on KAFKA-17174:
----------------------------------------
[~ChrisEgerton] thanks for your response. I'd like to keep the discussion here
to avoid noise on https://github.com/apache/kafka/pull/16599
{quote}
that's been created and possibly started out-of-band, which would actually be
less convenient than our current embedded cluster API and IMO is not a path we
should pursue.
{quote}
`ClusterInstance` can be started in-band if you disable the auto-start.
{quote}
TL;DR: While I'm also not opposed to adopting the ClusterTestExtensions API in
Connect, I don't think we should start work on that until we have a plan in
place for how we can do it for all of our integration tests instead of just
some.
{quote}
You are right that we should have a plan for it. Maybe we can start by
`EmbeddedKafkaCluster`:
1. change `EmbeddedKafkaCluster` to interfact
2. offer two impls of `EmbeddedKafkaCluster`. origin impl + impl by
`ClusterInstance`.
3. `EmbeddedConnect` will have another constructor which can take
`ClusterInstance` as argument
Also, we can enhance `ClusterInstance` according to use case of connect in this
migration.
{quote}
I don't think we should fragment Connect's integration tests and use
EmbeddedKafkaCluster in some but ClusterInstance in others
{quote}
Unfortunately, fragment must happen since connect is a huge module. It is hard
to address migration in few PRs. We have similar experience in migrating mock
tool and updating junit, right
> Migrate OffsetsApiIntegrationTest to use ClusterTestExtensions
> --------------------------------------------------------------
>
> Key: KAFKA-17174
> URL: https://issues.apache.org/jira/browse/KAFKA-17174
> Project: Kafka
> Issue Type: Improvement
> Reporter: Chia-Ping Tsai
> Assignee: Chia-Ping Tsai
> Priority: Minor
>
> Connect module has its `EmbeddedKafkaCluster`, so another alternative is to
> re-implement it by `ClusterInstance`. However, both of them are a kind of
> "embedded service" interface. Personally, I prefer to unify the interface for
> all modules in testing. We can move any useful helpers from
> `EmbeddedKafkaCluster` to `ClusterInstance` in order to make other modules
> get benefits too.
> `OffsetsApiIntegrationTest` is a lightweight user of `EmbeddedKafkaCluster`,
> so it can be a good entrypoint to introduce `ClusterInstance` to connect
> module
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
