sync-diff-inspector: enforce splitter-strategy + add auto mode

[joechenrh]

What problem does this PR solve?

Issue Number: close #xxxx

Today splitter-strategy in sync-diff-inspector is consumed only as a fallback when bucket-based splitting fails. The bucket iterator is always attempted first for chunk checksum, and global checksum ignores the setting entirely and always uses random. When TiDB bucket statistics are stale or misleading, bucket-based splits can produce skewed or incorrect chunk boundaries — and the user has no way to opt out.

What is changed and how it works?

splitter-strategy is now a three-valued enforced setting:

Value	Behavior
auto (new default)	Chunk checksum: try bucket first, fall back to random on error. Global checksum: random. Preserves the pre-existing behavior.
random	Random iterator for both chunk checksum and global checksum. Bucket iterator is not attempted.
limit	Limit iterator for both chunk checksum and global checksum. Bucket iterator is not attempted.

Implementation:

• sync_diff_inspector/config/config.go: add SplitterStrategyAuto and update normalizeSplitterStrategy so empty/"auto" normalize to auto, and any other unknown value fails CheckConfig.
• sync_diff_inspector/source/tidb.go:AnalyzeSplitter: switch on SplitterStrategy up-front. limit/random build their iterators directly; only auto falls through to the bucket-first path.
• sync_diff_inspector/source/tidb.go:GetGlobalChecksumIterator: switch on SplitterStrategy — limit returns a *LimitIterator (chunk count reported as 0 since LimitIterator produces chunks asynchronously with no upfront total); auto/random continue to use the random iterator.

Check List

Tests

• Unit test

New / updated unit tests:

• TestNormalizeSplitterStrategy in config/config_test.go — covers empty, auto, mixed case, surrounding whitespace, valid enum values, and an invalid value.
• TestAnalyzeSplitterEnforcesLimitStrategy / TestAnalyzeSplitterEnforcesRandomStrategy / TestAnalyzeSplitterAutoFallsBackToRandomOnBucketError in source/tidb_splitter_test.go — verify (via sqlmock + iterator type assertions) that non-auto strategies skip the bucket path, and that auto still falls back to random on bucket failure.
• TestGetGlobalChecksumIteratorUsesLimitWhenConfigured / TestGetGlobalChecksumIteratorDefaultsToRandom in source/tidb_checksum_test.go — verify the global-checksum dispatch honors limit and treats auto/random/"" identically as random.

Questions

Will it cause performance regression or break compatibility?

No performance regression. The default (auto) reproduces today's behavior byte-for-byte. Users who previously set splitter-strategy = "random" will now skip the bucket iterator — that matches the intent of having set the field explicitly.

Do you need to update user documentation, design documentation or monitoring documentation?

No. splitter-strategy is not publicly documented.

Release note

sync-diff-inspector: `splitter-strategy` is now enforced. Setting `random` or `limit` skips the bucket iterator entirely for both chunk checksum and global checksum. A new `auto` value (the default) preserves the previous bucket-first-with-random-fallback behavior.

[ti-chi-bot]

[FORMAT CHECKER NOTIFICATION]

Notice: To remove the do-not-merge/needs-linked-issue label, please provide the linked issue number on one line in the PR body, for example: Issue Number: close #123 or Issue Number: ref #456.

_{📖 For more info, you can check the "Contribute Code" section in the development guide.}

[ti-chi-bot]

[APPROVALNOTIFIER] This PR is NOT APPROVED

This pull-request has been approved by:
Once this PR has been reviewed and has the lgtm label, please assign overvenus for approval. For more information see the Code Review Process.
Please ensure that each of them provides their approval before proceeding.

The full list of commands accepted by this bot can be found here.

Details

Needs approval from an approver in each of these files:

• OWNERS

Approvers can indicate their approval by writing /approve in a comment
Approvers can cancel approval by writing /approve cancel in a comment

[gemini-code-assist]

Code Review

This pull request introduces a new "auto" splitter strategy for the sync_diff_inspector, making it the default option. The changes include updating the configuration normalization logic, modifying the TiDB table analyzer to explicitly handle "limit", "random", and "auto" strategies, and adding support for the "limit" strategy in global checksum iterations. Additionally, new unit tests were implemented to ensure correct strategy enforcement and configuration normalization. I have no feedback to provide as the existing review comments did not identify actionable issues or improvements within the scope of the provided code.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 53.4152%. Comparing base (7e0d4db) to head (55323a0).
⚠️ Report is 1 commits behind head on master.
✅ All tests successful. No failed tests found.

❌ Your project check has failed because the head coverage (53.4152%) is below the target coverage (60.0000%). You can increase the head coverage or adjust the target coverage.

Additional details and impacted files

Components	Coverage Δ
cdc	57.3566% <ø> (-0.0234%)	⬇️
dm	49.1975% <ø> (+0.0112%)	⬆️
engine	50.6659% <ø> (-0.0678%)	⬇️

Flag	Coverage Δ
cdc	57.3566% <ø> (?)
unit	53.4152% <ø> (-0.1003%)	⬇️

Flags with carried forward coverage won't be shown. Click here to find out more.

@@               Coverage Diff                @@
##             master     #12609        +/-   ##
================================================
- Coverage   53.5155%   53.4152%   -0.1003%     
================================================
  Files          1037       1011        -26     
  Lines        145625     139975      -5650     
================================================
- Hits          77932      74768      -3164     
+ Misses        61845      59588      -2257     
+ Partials       5848       5619       -229     

🚀 New features to boost your workflow:

• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.

[ti-chi-bot]

@joechenrh: The following tests failed, say /retest to rerun all failed tests or /retest-required to rerun all mandatory failed tests:

Test name	Commit	Details	Required	Rerun command
pull-cdc-integration-kafka-test	55323a0	link	true	/test pull-cdc-integration-kafka-test
pull-dm-integration-test-next-gen	55323a0	link	false	/test pull-dm-integration-test-next-gen
pull-dm-integration-test	55323a0	link	true	/test pull-dm-integration-test

Full PR test history. Your PR dashboard.

Details

Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes-sigs/prow repository. I understand the commands that are listed here.