Only run sidecardb change detection on serving primary tablets

[GuptaManan100]

Description

We noticed that the sidecardb logic to detect schema changes in vitess internal tables is running on both the transitions of going to primary serving and primary non-serving. We think it would be a good idea to only do this when a primary is transitioning to serving state. A couple of reasons for it -

1. DemotePrimary is already quite a heavy operation that sometimes times out, so we shouldn't do more work on this call.
2. If a primary tablet already applied the DDL changes when it went into serving state, then there should be no DDLs pending to be applied when it demoting itself. The check for finding the schema diff is therefore wasted effort at that point.
3. In EmergencyReparentShard, we demote the primary in parallel with stopping replication on replicas. This means that if even if there were to happen a DDL, the query could just theoretically block on semi-sync (it is a race), and that would fail DemotePrimary too.

This PR makes the change of passing in the desired serving state that we are transitioning to and makes the sidecardb code only run for serving primary transitions.

Related Issue(s)

• Fixes Bug Report: We run schema diff queries and changes even when a primary is going non-serving. #17060

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

All modified and coverable lines are covered by tests ✅

Project coverage is 67.15%. Comparing base (17607fa) to head (2ad880e).
Report is 3 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main   #17051      +/-   ##
==========================================
+ Coverage   67.14%   67.15%   +0.01%     
==========================================
  Files        1571     1571              
  Lines      252060   252061       +1     
==========================================
+ Hits       169249   169282      +33     
+ Misses      82811    82779      -32     

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

[arthurschreiber]

I agree this is a sensible change. 👍

3. In EmergencyReparentShard, we demote the primary in parallel with stopping replication on replicas. This means that if even if there were to happen a DDL, the query could just theoretically block on semi-sync (it is a race), and that would fail DemotePrimary too.

Does this qualify this change for backporting to older releases?

[shlomi-noach]

This absolutely must only ever run on a Primary.

[GuptaManan100]

Does this qualify this change for backporting to older releases?

I don't think so. During ERS, we already have writes incoming from the user too, which can potentially block on semi-sync. So, I think that problem is inherently there. I think it's a good idea to prevent DDLs like these too from blocking too, but I don't think it's serious enough to warrant backports.