CVE-2025-21522

What it actually does

The parser component in MySQL Server allows a low-privileged network attacker to trigger a hang or repeatable crash, resulting in complete denial of service for the database process.

Why it matters to the business

Database services become unavailable, causing application downtime, interrupted transactions, and potential loss of revenue or operational continuity until the server is restarted.

How an attack unfolds

An attacker with valid low-privilege database credentials connects over supported protocols and submits specially crafted statements that exploit a flaw in the parser, forcing the server into a hung or crashed state without requiring elevated rights or user interaction.

How to fix it

Upgrade affected MySQL Server installations to versions later than 8.0.40, 8.4.3, or 9.1.0 as provided by Oracle; apply the corresponding fixed packages for any downstream distributions that include MySQL components.

How to confirm the fix worked

Confirm the running MySQL version is newer than the listed vulnerable releases, review release notes for parser fixes, and test that repeated malformed queries no longer cause process termination.

Typical attack path

1. Obtain low-privileged network access to a vulnerable MySQL instance
2. Establish a connection using any supported protocol
3. Submit crafted input that reaches the parser component
4. Trigger repeated crashes or hangs to sustain denial of service

How to detect it

Review MySQL error logs and general query logs for repeated parser failures, abrupt process terminations, or connection resets originating from the same low-privileged accounts; correlate with network connection attempts around the time of availability loss.

What your security team should do

• Prioritize patching all internet-facing or multi-user MySQL instances within the next maintenance window
• Restrict network access to the database to only trusted clients and enforce least-privilege accounts
• Monitor for unusual connection volumes or repeated parser-related error patterns in database logs