DDL: do not Reload() for 'CREATE TEMPORARY' and 'DROP TEMPORARY' statements

[shlomi-noach]

Description

In query_executor.go, in execDDL(), we differentiate two cases:

1. The query is an Online DDL
2. The query is not Online DDL

In the latter case, we always invoke Reload() (specifically ReloadAtEx()) to reload table schema.
However, for CREATE TEMPORARY TABLE and DROP TEMPORARY TABLE this is wasteful:

• Only the connections which created or dropped the table is at all aware of the existence of the table
• Reload() adds an overhead which becomes significant if temporary tables are created and dropped intensively.

With this PR we skip Reload() for temporary tables.

Related Issue(s)

Checklist

• "Backport to:" labels have been added if this change should be back-ported
• Tests were added or are not required
• 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.
• If this is a change that users need to know about, please apply the release notes (needs details) label so that merging is blocked unless the summary release notes document is included.
• If a test is added or modified, there should be a documentation on top of the test to explain what the expected behavior is what the test does.

If a new flag is being introduced:

• Is it really necessary to add this flag?
• Flag names should be clear and intuitive (as far as possible)
• Help text should be descriptive.
• Flag names should use dashes (-) as word separators rather than underscores (_).

If a workflow is added or modified:

• Each item in Jobs should be named in order to mark it as required.
• If the workflow should be required, the maintainer team should be notified.

Bug fixes

• There should be at least one unit or end-to-end test.
• The Pull Request description should include a link to an issue that describes the bug.

Non-trivial changes

• There should be some code comments as to why things are implemented the way they are.

New/Existing features

• Should be documented, either by modifying the existing documentation or creating new documentation.
• New features should have a link to a feature request issue or an RFC that documents the use cases, corner cases and test cases.

Backward compatibility

• Protobuf changes should be wire-compatible.
• Changes to _vt tables and RPCs need to be backward compatible.
• vtctl command output order should be stable and awk-able.
• RPC changes should be compatible with vitess-operator
• If a flag is removed, then it should also be removed from VTop, if used there.

[harshit-gangal]

In the analyzeDDL plan
we mark in the plan NeedsReservedConn to true, so we can use that field here and ignore the reload

return &Plan{PlanID: PlanDDL, FullQuery: fullQuery, FullStmt: stmt, NeedsReservedConn: stmt.IsTemporary()}, nil

[shlomi-noach]

Can there be a future scenario where NeedsReservedConn is true even if the table is not TEMPORARY?

[shlomi-noach]

I've meanwhile simplified the check for IsTemporary()

[harshit-gangal]

I do not expect that to change and I have added a test specifically for that.
We can keep the current simplified version as well.

[shlomi-noach]

let's keep the existing version, then, as it is more explicit 🙏

[harshit-gangal]

Can we add a test for this change?

[shlomi-noach]

@harshit-gangal good idea; I don;t have an immediate understanding of how to test this, I'd need to dig in. If you happen to know, would you add such test?

[shlomi-noach]

Fixes #12146

[harshit-gangal]

@harshit-gangal good idea; I don;t have an immediate understanding of how to test this, I'd need to dig in. If you happen to know, would you add such test?

I will add the test

[shlomi-noach]

Tests-wise, I see validation that NeedsReservedConn is true or is false; but should there be a test that validates that Reload does, in fact, execute or does not execute?

[harshit-gangal]

Tests-wise, I see validation that NeedsReservedConn is true or is false; but should there be a test that validates that Reload does, in fact, execute or does not execute?

I am adding a reload time metric that will help achieve that test.