Implement server side cursors

[DvdGiessen]

Description

To be able to use the MySQL foreign data wrapper for PostgreSQL, we have to support server side cursors.

To that end each connection now keeps track of a list of open cursors, and implements the COM_STMT_FETCH command for retrieving data from an open cursor.

Based on the implementation in the Dolthub fork of Vitess:

• simulate server side cursors dolthub/vitess#228
• go/mysql/conn.go: Improve server-side cursor support. dolthub/vitess#232

Compared to the implementation in the Dolthub fork this commit:

• Was changed to fit with all changes made in Vitess since the Dolthub version was forked
• Remove a few superfluous formatting/whitespace changes made in there to make reviewing easier.
• Also supports multiple open cursors whereas the Dolthub fork only supports a single open cursor.

I've also moved all the changes to other code (namely, updating them to use the "status" integer instead of the old "more" boolean) into a separate commit to make it easier to review.

Related Issue(s)

#18046

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

This only implements a new protocol feature and does not affect any on-disk information. No migrations should be 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.

[DvdGiessen]

Re-pushed with updated commit description because I just noticed I misattributed some of the original work this change is based on (I copied the author from a wrong commit). Want to make sure I give credit where it's due. (:

[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]

You would need to add e2e test to validate it works cleanly with fdw

[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]

There are some CI failures that needs to be addressed.

[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.

[DvdGiessen]

You would need to add e2e test to validate it works cleanly with fdw

I agree a E2E test for this would be useful to prevent this from breaking in the future.

However, I'm not entirely sure how to go about implementing this E2E test within the template structure used for generating the workflows. I'd very much welcome if anyone else wants to help contribute such a test!

The basic test would work like this:

• Create some tables in Vitess which we want to test querying.

• Start a PostgreSQL server with the mysql_fdw extension installed.

• Run SQL on PostgreSQL to set up:

  CREATE DATABASE vitess;
  \c vitess
  CREATE EXTENSION mysql_fdw;
  CREATE SERVER vitess FOREIGN DATA WRAPPER mysql_fdw OPTIONS (
      host 'vtgate',
      port '3306',
      sql_mode 'STRICT_TRANS_TABLES',
      use_remote_estimate 'true'
  );
  CREATE USER MAPPING FOR postgres SERVER vitess OPTIONS (
      username 'root',
      password 'qwerty123'
  );
  IMPORT FOREIGN SCHEMA postgres_fdw FROM SERVER vitess INTO public;

• Do some test queries? For example:

  • A simple unsharded select
  • A big shared select
  • A select with pushdown from PostgreSQL to Vitess (a where clause, a join)

Also note that while compatibility with PostgreSQL FDW happens to be my use case, there is nothing specific to PostgreSQL in the changes made in this PR. It implements support for cursors in the MySQL protocol, and thus should be useful to anyone that wants to use cursors for any reason.

Presumably there could also be E2E tests that simply use a cursor from a regular MySQL client. I think having a specific E2E test for PostgreSQL FDW, while definitely a nice to have, doesn't have to be a specific blocker for this PR.

[harshit-gangal]

simply use a cursor from a regular MySQL client

This is exactly what I am looking for in our e2e test. we do not want postgres running in our CI

[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.

[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.