Allow multi-shard read-only transactions in SINGLE transaction mode

[harshit-gangal]

Description

This change updates the behavior of transaction_mode = SINGLE to allow transactions that span multiple shards, as long as they only execute read queries.

Previously, any query that opened sessions on more than one shard would cause the transaction to fail, even if all operations were reads. With this change:

• A transaction is allowed to touch multiple shards if all of them are read-only.
• The transaction will only fail if more than one shard has executed a write (i.e., data-modifying queries).

This avoids unnecessary failures in common multi-shard read use cases, while still enforcing the single-shard write guarantee of SINGLE mode.

Related Issue(s)

• Closes Feature Request: Allow Transaction Mode Single to Execute Read only queries to multiple shards #18285

Checklist

• 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

[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 83.33333% with 3 lines in your changes missing coverage. Please review.

Project coverage is 67.47%. Comparing base (c3daef3) to head (9f44fd9).
Report is 21 commits behind head on main.

Files with missing lines	Patch %	Lines
go/vt/vtgate/executorcontext/vcursor_impl.go	40.00%	3 Missing ⚠️

Additional details and impacted files

@@           Coverage Diff           @@
##             main   #18173   +/-   ##
=======================================
  Coverage   67.46%   67.47%           
=======================================
  Files        1602     1602           
  Lines      262245   262243    -2     
=======================================
+ Hits       176930   176946   +16     
+ Misses      85315    85297   -18     

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

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.

[eeSeeGee]

It'd be nice to have a way to test for queries that would bounce against the new policy rather than actually bouncing it (producing a metric). That'd give us confidence to roll it out in prod

[aparajon]

Would it be possible to make cross shard writes opt-in via a comment directive (example)?

[github-actions]

This PR is being marked as stale because it has been open for 30 days with no activity. To rectify, you may do any of the following:

• Push additional commits to the associated branch.
• Remove the stale label.
• Add a comment indicating why it is not stale.

If no action is taken within 7 days, this PR will be closed.

[harshit-gangal]

It'd be nice to have a way to test for queries that would bounce against the new policy rather than actually bouncing it (producing a metric). That'd give us confidence to roll it out in prod

@eeSeeGee There is no easy way to add this metric. What mode currently you are running in production?

[GrahamCampbell]

Isn't this flawed if the select executes a sproc or UDF that does writes?

[harshit-gangal]

Would it be possible to make cross shard writes opt-in via a comment directive (example)?

It is totally doable to add transaction_mode to comment directive.

I can add that in a follow up PR

[harshit-gangal]

Isn't this flawed if the select executes a sproc or UDF that does writes?

@GrahamCampbell Only Select/Show statement are considered read-only and others are considered as write.
sproc is invoked with a call which will be considered a write op.
AFAIK UDFs does not allow modification of data.

[aparajon]

Would it be possible to make cross shard writes opt-in via a comment directive (example)?

It is totally doable to add transaction_mode to comment directive.

I can add that in a follow up PR

That's amazing @harshit-gangal, thank you! I can foresee us heavily using that feature at Block to guardrail and test queries