MySQL 8.0 - Unsupported Redo Log Format
Upgrading a production database or even a local XAMPP instance from MySQL 5.x to version 8.0 is often more than a simple package replacement. The jump between these versions introduces architectural changes that the server detects immediately upon startup, specifically regarding how InnoDB manages its transaction logs. If you encounter an error stating Unsupported redo log format (0) in your logs, it means MySQL 8.0 refuses to start because it cannot safely interpret the existing binary log files left over from a previous, incompatible version.
The core issue lies in the transition of the redo log structure. Prior to version 5.7.9, InnoDB used a legacy redo log format. Starting with 5.7.9, MySQL adopted a new, more efficient format that significantly reduced the size of the logs and improved write performance. When you upgrade directly from an older version (like 5.6 or early 5.7) to 8.0 without clearing these files, the server reads the header of the ib_logfile files. It sees a format identifier indicating it was created before 5.7.9. Because MySQL 8.0 requires the newer format for its internal operations, including the Data Dictionary initialization, it aborts the plugin start process entirely to prevent data corruption or silent inconsistencies.
The error message Failed to initialize DD Storage Engine and Data Dictionary initialization failed confirms that the server cannot even boot up to the point where it could attempt a repair. The system is protecting itself by shutting down immediately rather than risking an upgrade of the log files in-place, which is technically unsupported and risky.
To resolve this, you must manually truncate or delete the old redo log files before starting the upgraded instance. This forces MySQL 8.0 to generate new log files with the correct format headers upon its first successful startup.
Identifying and Removing Old Log Files
The files causing the conflict are typically located in your data directory. In a standard XAMPP installation on Windows, this is usually found at C:\xampp\mysql\data. Look for files named ib_logfile0 and ib_logfile1. These are the redo log files referenced by the error.
You should not simply rename them to .old unless you have a specific backup strategy; deletion is the standard procedure for this upgrade scenario because the data inside these logs will be overwritten as soon as the server writes new transactions. However, ensure your innodb_log_file_size configuration in your my.ini file matches the expected size for version 8.0 if you intend to keep them, though usually, deleting them and letting MySQL recreate default-sized logs is safer during an upgrade path.
Execute the following steps carefully:
- Stop the MySQL service completely to ensure no files are locked by the OS.
- Navigate to your data directory.
- Delete
ib_logfile0andib_logfile1. - Optionally, check for
ib_logfile2if you previously had a three-file configuration, though two is standard.
Configuration Cleanup
While addressing the redo logs, it is prudent to review your my.ini file for deprecated settings that might have been valid in 5.x but are no longer needed or could cause warnings in 8.0. Common culprits include innodb_flush_log_at_trx_commit (which remains supported but has different performance characteristics) or specific buffer pool settings that changed behavior.
However, the primary blocker is the physical presence of the old log files. Once deleted, start the MySQL service again. The server will detect the missing files and automatically create new ones with the version-specific format required by 8.0. The startup logs should then show a clean initialization without the MY-013090 error.
Why Not Just Update In-Place?
The temptation might be to update MySQL while keeping the old files, hoping the server will "fix" them. This is not how the upgrade path works for InnoDB. The binary format of the redo logs is not version-neutral in a way that allows seamless upgrades across major versions like 5.6 to 8.0. The Data Dictionary (DD), which manages metadata about your tables, also relies on these logs being consistent with its own storage engine expectations. If the log format is old, the DD cannot verify transaction integrity, leading to the generic errors seen in the stack trace.
By clearing the redo logs, you are essentially performing a "clean slate" for the transaction logging subsystem while preserving your actual table data files (.ibd files) and configuration. This is a safe operation as long as you have your database dumps or backups ready, which you should always do before touching production data directories.
For developers managing legacy environments or upgrading local XAMPP stacks for testing, understanding this limitation prevents hours of debugging startup failures that are actually just file management issues. If you are deploying Laravel applications or other modern PHP stacks on MySQL 8.0, ensuring the upgrade path includes a clean redo log initialization is part of the standard best practice for version jumps beyond 5.7.9.