Add no-parallel-replicas tag to 02845_join_on_cond_sparse#101023
Draft
alexey-milovidov wants to merge 9 commits intomasterfrom
Draft
Add no-parallel-replicas tag to 02845_join_on_cond_sparse#101023alexey-milovidov wants to merge 9 commits intomasterfrom
no-parallel-replicas tag to 02845_join_on_cond_sparse#101023alexey-milovidov wants to merge 9 commits intomasterfrom
Conversation
The test depends on exact INSERT visibility: the second SELECT must see the freshly inserted row `(0, 1)`. Under parallel replicas the reading is distributed across replicas, and the fresh part may not be included in the coordinator snapshot, causing the SELECT to return empty. Found in: https://s3.amazonaws.com/clickhouse-test-reports/json.html?PR=100573&sha=540a0ab7b9db7f85d22143f2857ba15d2bd93304&name_0=PR&name_1=Stateless%20tests%20%28amd_tsan%2C%20parallel%2C%202%2F2%29 Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Contributor
|
Workflow [PR], commit [f5d3d55] Summary: ❌
AI ReviewSummaryThis PR updates ClickHouse Rules
Final Verdict
|
Member
Author
|
I also found this strange. Would you mind telling exactly why it didn't work with parallel replicas? |
…parse` The flaky check showed that this test also fails with randomized MergeTree settings (e.g. `min_bytes_for_wide_part 0`, large `index_granularity`, `enable_index_granularity_compression`, etc.) even without parallel replicas. Since this is a sparse serialization regression test that depends on specific table-level settings, prevent randomized MergeTree settings from interfering. https://s3.amazonaws.com/clickhouse-test-reports/json.html?PR=101023&sha=519b560776151364a61af63d857ff479ecc8f0c0&name_0=PR&name_1=Stateless%20tests%20%28amd_msan%2C%20flaky%20check%29 Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
| @@ -1,3 +1,5 @@ | |||
| -- Tags: no-parallel-replicas, no-random-merge-tree-settings | |||
Contributor
There was a problem hiding this comment.
The new -- Tags line also adds no-random-merge-tree-settings, but the PR description only justifies no-parallel-replicas.
Please either:
- drop
no-random-merge-tree-settingsto keep test coverage broader, or - add a concrete failure mode showing why random
MergeTreesettings make this test nondeterministic.
We try to add no-* tags only when they are strictly necessary.
The random MergeTree settings injector in `ClientBase::addMergeTreeSettings` skips settings already present in the CREATE TABLE statement, so the test's explicit `ratio_of_defaults_for_sparse_serialization = 0.1` is not overridden. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
…rse` The flaky check CI proves this tag is needed: the random MergeTree setting `ratio_of_defaults_for_sparse_serialization = 1.0` overrides the table-level setting of `0.1`, disabling sparse serialization which this test specifically exercises. This causes the SELECT to return empty results. Evidence: - https://s3.amazonaws.com/clickhouse-test-reports/json.html?PR=101023&sha=be1a25bec699d8788154004531cdd5475d6bcae1&name_0=PR&name_1=Stateless%20tests%20%28amd_debug%2C%20flaky%20check%29 - https://s3.amazonaws.com/clickhouse-test-reports/json.html?PR=101023&sha=be1a25bec699d8788154004531cdd5475d6bcae1&name_0=PR&name_1=Stateless%20tests%20%28amd_binary%2C%20flaky%20check%29 Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
The table-level SETTINGS (`ratio_of_defaults_for_sparse_serialization`) are not overridden by random session-level MergeTree settings, so the tag is not strictly necessary. Keep only `no-parallel-replicas` which is justified by the INSERT visibility issue. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
The test depends on exact INSERT visibility: the second SELECT must see the freshly inserted row
(0, 1). Under parallel replicas the reading is distributed across replicas, and the fresh part may not be included in the coordinator snapshot, causing the SELECT to return empty.Found in: https://s3.amazonaws.com/clickhouse-test-reports/json.html?PR=100573&sha=540a0ab7b9db7f85d22143f2857ba15d2bd93304&name_0=PR&name_1=Stateless%20tests%20%28amd_tsan%2C%20parallel%2C%202%2F2%29
Changelog category (leave one):