VDiff: Save lastpk value for source and target

[mattlord]

Description

The RowStreamer accepts a lastpk value in order to start streaming from after a given row in the table (it uses the PrimaryKey columns to build the WHERE and ORDER BY clauses for the resultset that it streams).

VDiff passes in a lastpk value when resuming due to error or user request — for both the sources and targets of the workflow. When the PrimaryKey definition differs for a table between the source and target, however, this would fail if the lastpk value's fields did not match the table definition. So if the PK definition differed between the source and target the VDiff resume would fail with an error like this:

"vttablet: rpc error: code = Unknown desc = primary key values don't match length: [INT32(89687) INT64(5075599148)] vs [0]"

This PR fixes this issue by storing a lastpk value for both the source and target, as needed, and passing the respective value to the RowStreamer on the source and target tablets.

Related Issue(s)

• Fixes: Bug Report: VDiff fails when source and target have a different PK definition #17499

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

Attention: Patch coverage is 64.08451% with 51 lines in your changes missing coverage. Please review.

Project coverage is 67.70%. Comparing base (06def14) to head (b94eb40).
Report is 9 commits behind head on main.

Files with missing lines	Patch %	Lines
go/vt/vttablet/tabletmanager/vdiff/table_differ.go	64.91%	40 Missing ⚠️
...vt/vttablet/tabletmanager/vdiff/workflow_differ.go	65.38%	9 Missing ⚠️
.../vt/vttablet/tabletserver/vstreamer/rowstreamer.go	0.00%	2 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main   #17493      +/-   ##
==========================================
+ Coverage   67.69%   67.70%   +0.01%     
==========================================
  Files        1584     1584              
  Lines      254466   254597     +131     
==========================================
+ Hits       172258   172383     +125     
- Misses      82208    82214       +6     

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

[deepthi]

One of the things that made this issue hard to understand is the log message

primary key values don't match length: [INT32(89687) INT64(5075599148)] vs [0]

That is coming from

	if len(rs.lastpk) != 0 {
		if len(rs.lastpk) != len(rs.pkColumns) {
			return "", fmt.Errorf("primary key values don't match length: %v vs %v", rs.lastpk, rs.pkColumns)
		}

From this code snippet, pkColumns is a slice of length 1 with value 0. What it actually represents is the column index of the PK columns, not the actual PK column values. Is there any way to make this error message easier to understand? It would have been more obvious if it had said

primary key values don't match length: [INT32(89687) INT64(5075599148)] vs 1

though I understand that for debugging purposes, it might be helpful to log the entire pkColumns slice (as we are doing today) versus its length.