VReplication workflows: retry "wrong tablet type" errors

[rohit-nayak-ps]

Description

If a tablet picker is configured to pick REPLICA tablet types only, and when it starts streaming the selected tablet is promoted to a PRIMARY because of a PRS or ERS, the vttablet returns an grpc error like "error": "vttablet: rpc error: code = FailedPrecondition desc = wrong tablet type: PRIMARY, want: REPLICA or []",

Currently all FailedPreconditions are treated as unrecoverable errors for vreplication workflows. This PR adds an exception for the wrong tablet FailedPreconditionerror treats it as recoverable.

This will resolve situations where there was race where

• the tablet picker on the target chooses a tablet, type REPLICA
• the target is about to do a VStream RPC to that tablet. The RPC request has query.Target which has the expected TabletType
• there is PRS and the tablet promotes
• the tablet then responds to the RPC, but fails because the request's tablet type doesn't match the current tablet type

Note:
It is possible that this is not a race, but a missing watcher event. In that case, unless the local health check state gets refreshed with the new info, we will still see the same errors, if that tablet was selected again.

Backport reason: We should fix this on all supported versions, since it is a long-standing issue.

Related Issue(s)

#16646

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

[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 68.97%. Comparing base (81ed314) to head (7fbf290).
Report is 7 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main   #16645      +/-   ##
==========================================
+ Coverage   68.96%   68.97%   +0.01%     
==========================================
  Files        1562     1562              
  Lines      200730   200733       +3     
==========================================
+ Hits       138430   138457      +27     
+ Misses      62300    62276      -24     

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

[shlomi-noach]

This is a very simple solution and makes perfect sense.

I don't like how we identify the error by text/regexp; we should use something more formal, like a specific error object, or some error code, or something. But this is outside the cope of this PR.

[rohit-nayak-ps]

I don't like how we identify the error by text/regexp; we should use something more formal, like a specific error object, or some error code, or something. But this is outside the scope of this PR.

I reused how query serving and vttablet state manager were using that error in go/vt/vterrors/constants.go. But I agree, we should probably refactor that to use a typed error object.

[shlomi-noach]

I reused how query serving and vttablet state manager

Absolutely. This was a side note unrelated to the PR.

[deepthi]

LGTM

[shlomi-noach]