vtctld: fix GetTablets stale primary report under --tablet-alias filter

[Taeknology]

Description

GetTablets --tablet-alias was reporting stale former primaries as primary when the real current primary wasn't in the request. The fix consults the shard record via GetShard instead of trusting the result-set max, hybridised with the existing seed so transient cases (no recorded primary yet) stay best-effort.

The shard-filter path (--keyspace ... --shard ...) was correct already because the result set always includes the real primary. The alias-filter path doesn't — hence the bug.

Related Issue(s)

Fixes #19898

Reproducer tests by @mhamza15 (first commit, cherry-picked verbatim).

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

Backport candidates: release-24.0 and release-23.0 — wrong primary reports from --tablet-alias can confuse operator automation. Tag if you agree.

Deployment Notes

Stale primaries queried by --tablet-alias now report as unknown instead of primary. Review any monitoring that relied on the old behavior. Adds one parallel GetShard per unique shard in the result.

AI Disclosure

Most of this was written by Claude Code - I just provided direction.

[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.

[Copilot]

Pull request overview

Fixes a bug in vtctld GetTablets where filtering by --tablet-alias could misreport stale former primaries as primary, because the previous code computed the "true primary" only as the max PrimaryTermStartTime within the result set (which excludes the real current primary when only stale aliases are queried). The fix consults GetShard per unique shard to obtain the authoritative term start time, falling back to the existing seed when the shard record is zero, and parallelizes those lookups with strict/non-strict failure modes.

Changes:

• Replace the single truePrimaryTimestamp with a per-shard map seeded from result-set primaries, then overwrite each entry with Shard.PrimaryTermStartTime fetched via parallel GetShard calls.
• Skip stale-primary adjustment for shards whose GetShard failed; aggregate errors with errors.Join, returning them in strict mode and logging a warning otherwise.
• Add table-driven tests covering single-alias, multi-shard alias, all-current-primaries, and strict/non-strict GetShard failure scenarios, plumbing addTabletOptions and a factorySetup hook through the test harness.

Reviewed changes

Copilot reviewed 2 out of 2 changed files in this pull request and generated 1 comment.

File	Description
go/vt/vtctl/grpcvtctldserver/server.go	Per-shard primary term resolution via concurrent GetShard with strict/non-strict error handling.
go/vt/vtctl/grpcvtctldserver/server_test.go	New test cases for stale-primary alias filtering and topo-failure paths; harness now supports AddTabletOptions and topo factory hooks.

[Taeknology]

Status of Needs* labels

NeedsIssue — satisfied. PR body links Fixes #19898.

NeedsWebsiteDocsUpdate — satisfied by acabf849f8
(adds VTCtld minor-changes entry to
changelog/25.0/25.0.0/summary.md).

[Taeknology]

Backport feasibility (verified locally)

I cherry-picked the fix onto both candidate branches and ran the
tests:

• release-24.0: cherry-pick is clean (auto-merge handles a small
  import-block diff). All 5 new test cases and the full
  TestGetTablets pass with no regression.
• release-23.0: cherry-pick has two trivial 3-way merges (import
  block, and the test setup harness — t.Context() +
  NewServerAndFactory, both available on Go 1.25). One adaptation is
  required: the fix uses the slog-style
  log.Warn(..., slog.Any/Int(...)) API introduced in Add structured logging with slog #19327, which
  was backported to release-24.0 but not to release-23.0. Replacing
  the call with
  log.Warningf("...affected_shards=%d, error=%v", ...) and dropping
  the log/slog import builds cleanly. All 5 new tests and the full
  TestGetTablets pass with no regression.

If you agree, please add:

• Backport to: release-24.0
• Backport to: release-23.0

I'll handle the log.Warn adaptation on the draft backport PR the bot
opens for release-23.0.