ryankert01 opened a new issue, #4544:
URL: https://github.com/apache/iggy/issues/4544

   ### Description
   
   `.config/nextest.toml` [still serializes the Elasticsearch connector 
tests](https://github.com/apache/iggy/blob/bb1ef105e54483aea94864c694457f42688cdfa7/.config/nextest.toml#L30-L40).
 The comment above the group says this avoids a Docker 409 Conflict when 
parallel tests race to create the shared `iggy-test-elasticsearch` container.
   
   That race is already handled in the fixture. 
[`start_shared_container`](https://github.com/apache/iggy/blob/bb1ef105e54483aea94864c694457f42688cdfa7/core/integration/tests/connectors/fixtures/elasticsearch/container.rs#L126-L178)
 retries the 409 up to 30 times, and each retry attaches to the container the 
winner created. The [identical group for Doris was 
removed](https://github.com/apache/iggy/commit/d693eb93d878a7759221fba05dcea20e52dd1249)
 for the same reason, since its fixture uses the same design, and CI stayed 
green across all three partitions with no retries or timeouts. So the 
Elasticsearch group is redundant, and its comment gives a reason that is no 
longer true.
   
   While aligning the two fixtures, the Doris one has two small gaps compared 
with Elasticsearch:
   
   - It [sets no 
`with_startup_timeout`](https://github.com/apache/iggy/blob/bb1ef105e54483aea94864c694457f42688cdfa7/core/integration/tests/connectors/fixtures/doris/container.rs#L240-L254),
 so the test that creates the container gets testcontainers' [60s 
default](https://github.com/testcontainers/testcontainers-rs/blob/0.27.3/testcontainers/src/runners/async_runner.rs#L25)
 for the FE health wait, while [Elasticsearch uses 
120s](https://github.com/apache/iggy/blob/bb1ef105e54483aea94864c694457f42688cdfa7/core/integration/tests/connectors/fixtures/elasticsearch/container.rs#L147).
 Tests that attach [skip the 
wait](https://github.com/testcontainers/testcontainers-rs/blob/0.27.3/testcontainers/src/runners/async_runner.rs#L139-L159),
 so only the creator is affected.
   - [`check_vm_max_map_count` reports `fixture_type: 
"SharedDoris"`](https://github.com/apache/iggy/blob/bb1ef105e54483aea94864c694457f42688cdfa7/core/integration/tests/connectors/fixtures/doris/container.rs#L187),
 while every other Doris error uses `"DorisContainer"`.
   
   ### Affected area / component
   
   Connectors, CI / build / tooling
   
   ### Proposed solution
   
   - Delete `[test-groups.elasticsearch]` and its override from 
`.config/nextest.toml`.
   - Doris fixture: add `.with_startup_timeout(Duration::from_secs(120))` and 
rename `"SharedDoris"` to `"DorisContainer"`.
   
   The PR will include CI timings for the Elasticsearch tests run in parallel, 
as the Doris change did.
   
   ### Alternatives considered
   
   Keep the group and update its comment to give a different reason. I could 
not find one: per-test isolation already comes from a unique index per fixture.
   
   ### Contribution
   
   - [x] I'm willing to submit a pull request to implement this feature
   


-- 
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]

Reply via email to