This is an automated cherry-pick of #39847

What problem does this PR solve?

Issue Number: close #38950

Problem Summary:

What is changed and how it works?

The code below is added by PR #30301 to fix Issue #29483. When replace encounters a duplicate PRIMARY KEY or UNIQUE KEY, it is expected to have the semantics of DELETE + INSERT, when Auto ID is out of range, the truncated ID is possible to duplicate with an existing ID, this leads to data corruption as unrelated rows are modified and removed.

if err == nil && d.GetInt64() < id {
    // Auto ID is out of range, the truncated ID is possible to duplicate with an existing ID.
    // To prevent updating unrelated rows in the REPLACE statement, it is better to throw an error.
    return autoid.ErrAutoincReadFailed
}

So when the ignore keyword exists in the SQL (or SET sql_mode='NO_ENGINE_SUBSTITUTION';), if the Auto ID is out of range, the above code prevents replace into before it is actually executed, but it also prevents a lot SQLs that need to be executed, so special handling is required when an auto id out of range error occurs in insert ignore into ... on duplicate ... to allow SQL to be executed successfully.

Check List

Tests

• Unit test
• Integration test
• Manual test (add detailed scripts or steps below)
• No code

Side effects

• Performance regression: Consumes more CPU
• Performance regression: Consumes more Memory
• Breaking backward compatibility

Documentation

• Affects user behaviors
• Contains syntax changes
• Contains variable changes
• Contains experimental features
• Changes MySQL compatibility

Release note

Please refer to Release Notes Language Style Guide to write a quality release note.

None