schemadiff: support valid foreign key cycles

[shlomi-noach]

Description

Addressing #15430, in this PR schemadiff uses the same algorithm as Query-Serving to distinguish between valid and invalid foreign key cycles. schemadiff now allows table cycles as long as there's no column-reference cycle.

There's a few implications to this solution. Consider this valid foreign key cycle:

create table t11 (id int primary key, i int, constraint f11 foreign key (i) references t12 (id));
create table t12 (id int primary key, i int, constraint f12 foreign key (i) references t11 (id));

• There is technically no valid ordering to creating these two tables. In MySQL, this could only be achieved by setting FOREIGN_KEY_CHECKS=0. However, schemadiff is declarative. We introduce DiffHints.ForeignKeyCheckStrategy which can be:

  • ForeignKeyCheckStrategyStrict (default value), makes Diff(), OrderedDiffs() etc. behave as if FOREIGN_KEY_CHECKS=1
  • ForeignKeyCheckStrategyIgnore, assume behavior of FOREIGN_KEY_CHECKS=0.
    This means OrderedDiffs() can generate an invalid schema whilst applying changes. We add one final validation to ensure that the grand total result is valid.

• The ordering of tables in a Schema object is normally by:

  • tables, foreign key dependency order (parents first, children tables)
  • tables, lexicographically
  • views, dependency order
  • views, lexicographically

  (See related discussion in WIP: MoveTables Cancel: sort table list taking into account FK order and drop tables in reverse order. #15252 (comment)).
  However, with cyclic foreign key references the ordering of tables that either participate in a loop, or reference a loop, is undefined. It will always be higher than the order of tables that do not participate in a loop, but no deterministic ordering in between loop-related tables.

• It is up to the consumer of schemadiff to actually SET FOREIGN_KEY_CHECKS=0 in MySQL. schemadiff does not generate such SQL statement.

Related Issue(s)

• Feature Request: schemadiff & OnlineDDL: support valid foreign key table cycles #15430
• Prior work: schemadiff: analyze and report foreign key loops/cycles #15062

Checklist

• "Backport to:" labels have been added if this change should be back-ported to release branches
• If this change is to be back-ported to previous releases, a justification is included in the PR description
• Tests were added or are not required
• Did the new or modified tests pass consistently locally and on CI?
• Documentation was added or is not required

Deployment Notes

[vitess-bot]

Review Checklist

Hello reviewers! 👋 Please follow this checklist when reviewing this Pull Request.

General

• Ensure that the Pull Request has a descriptive title.
• Ensure there is a link to an issue (except for internal cleanup and flaky test fixes), new features should have an RFC that documents use cases and test cases.

Tests

• Bug fixes should have at least one unit or end-to-end test, enhancement and new features should have a sufficient number of tests.

Documentation

• Apply the release notes (needs details) label if users need to know about this change.
• New features should be documented.
• There should be some code comments as to why things are implemented the way they are.
• There should be a comment at the top of each new or modified test to explain what the test does.

New flags

• Is this flag really necessary?
• Flag names must be clear and intuitive, use dashes (-), and have a clear help text.

If a workflow is added or modified:

• Each item in Jobs should be named in order to mark it as required.
• If the workflow needs to be marked as required, the maintainer team must be notified.

Backward compatibility

• Protobuf changes should be wire-compatible.
• Changes to _vt tables and RPCs need to be backward compatible.
• RPC changes should be compatible with vitess-operator
• If a flag is removed, then it should also be removed from vitess-operator and arewefastyet, if used there.
• vtctl command output order should be stable and awk-able.

[codecov]

Codecov Report

Attention: Patch coverage is 95.45455% with 5 lines in your changes are missing coverage. Please review.

Project coverage is 65.66%. Comparing base (6c73053) to head (d3b8c8e).
Report is 7 commits behind head on main.

Files	Patch %	Lines
go/vt/schemadiff/schema_diff.go	70.00%	3 Missing ⚠️
go/vt/schemadiff/schema.go	96.07%	2 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main   #15431      +/-   ##
==========================================
- Coverage   65.72%   65.66%   -0.07%     
==========================================
  Files        1563     1563              
  Lines      194027   194380     +353     
==========================================
+ Hits       127529   127644     +115     
- Misses      66498    66736     +238     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[dbussink]

@shlomi-noach Should we change ForeignKeyLoopError to have a different Loop type? So something like this maybe:

type ForeignKeyColumn struct {
	Table  string
	Column string
}

type ForeignKeyLoopError struct {
	Table string
	Loop  []ForeignKeyColumn
}

This because we now better track this at a column level, so this way we have the details of what the loop looks like?

[shlomi-noach]

Good point. I'll do so.

[shlomi-noach]

Done in d3b8c8e