[TEST] Cover custom sorting and routing in randomized testing #120584
Conversation
kkrik-es
added
>test
| var responseBody = entityAsMap(response); | ||
| assertThat("errors in baseline bulk response:\n " + responseBody, responseBody.get("errors"), equalTo(false)); | ||
| return responseBody; | ||
| } |
There was a problem hiding this comment.
All this logic feels out of place here, specific to tests using bulk indexing.. It probably belongs more to a sibling of ReindexChallengeRestIT, like BulkChallengeRestIT. Thoughts?
There was a problem hiding this comment.
Went ahead with this refactoring, ptal.
… logsdb/random-test-routing
|
Pinging @elastic/es-storage-engine (Team:StorageEngine) |
|
I'll wait for this to land before merging #119994, i think it would be a much easier conflict resolve for me. |
| * mapping and documents in order to cover more code paths and permutations. | ||
| */ | ||
| public class StandardVersusLogsIndexModeRandomDataChallengeRestIT extends StandardVersusLogsIndexModeChallengeRestIT { | ||
| public abstract class StandardVersusLogsIndexModeRandomDataChallengeRestIT extends StandardVersusLogsIndexModeChallengeRestIT { |
There was a problem hiding this comment.
Doesn't this effectively remove all tests that were in this class (inherited from StandardVersusLogsIndexModeChallengeRestIT)? Those are the actual logsdb vs standard random data tests.
There was a problem hiding this comment.
These now run in BulkChallengeRestIT that inherits from StandardVersusLogsIndexModeRandomDataChallengeRestIT.
|
I have to admit i am lost in the class hierarchy here (and i majorly contributed to that) which i am not a fan of. I don't have any specific suggestion though so LGTM modulo the question above. |
It definitely feels it's deeper that it should, for a testing suite.. Here is the class hierarchy: 
I think it's good that we have symmetry between |
|
Can we collapse |
Not really, lemme do it and submit. |
… logsdb/random-test-routing
|
Alright moved tests with static mappings to be a leaf: 
|
# Conflicts: # x-pack/plugin/logsdb/src/javaRestTest/java/org/elasticsearch/xpack/logsdb/qa/StandardVersusLogsIndexModeChallengeRestIT.java
kkrik-es
added
the
auto-merge-without-approval
kkrik-es
removed
the
auto-merge-without-approval
…c#120584) * Cover custom sorting and routing in randomized testing * [CI] Auto commit changes from spotless * fix reindex tests * fix reindex tests * refactor classes * comment * more refactoring * more refactoring * restore tests with static mappings * reduce diff * reduce diff * Restore single-element array removal in synthetic source * Revert "Restore single-element array removal in synthetic source" This reverts commit e8e99e1. * [CI] Auto commit changes from spotless --------- Co-authored-by: elasticsearchmachine <[email protected]>
💚 Backport successful
|
… (#120932) * Cover custom sorting and routing in randomized testing * [CI] Auto commit changes from spotless * fix reindex tests * fix reindex tests * refactor classes * comment * more refactoring * more refactoring * restore tests with static mappings * reduce diff * reduce diff * Restore single-element array removal in synthetic source * Revert "Restore single-element array removal in synthetic source" This reverts commit e8e99e1. * [CI] Auto commit changes from spotless --------- Co-authored-by: elasticsearchmachine <[email protected]>
3 participants
Requires some refactoring due to the fact that, with routing on sort fields, ids are auto-generated. Since some tests compare base and contender results using doc ids, the auto-generated id from contender needs to be used when indexing docs on baseline side.
Related to #109334