MySQL/MariaDB Won't Start

Diagnose and fix MySQL or MariaDB that fails to start, common causes and step-by-step solutions

MySQL or MariaDB failing to start is usually caused by a handful of well-known issues: corrupt InnoDB logs, full disk, wrong permissions, or a port conflict. This guide covers the most common causes.

Step 1: Check the error logs

Always start here:

# MySQL
sudo journalctl -u mysql -n 100 --no-pager

# MariaDB
sudo journalctl -u mariadb -n 100 --no-pager

# Or check the error log file directly
sudo tail -50 /var/log/mysql/error.log

The error message in the log tells you exactly what's wrong in most cases.

Cause 1: Disk full

[ERROR] InnoDB: Write to file ./ibdata1 failed
[ERROR] Error writing file '/var/lib/mysql/binlog.000001'

Check disk usage:

df -h
du -sh /var/lib/mysql/

Free up space:

# Remove old binary logs (safe if no replicas depend on them)
sudo mysql -e "PURGE BINARY LOGS BEFORE NOW();"

# Or truncate logs
sudo journalctl --vacuum-size=200M

# Check for large temp files
find /tmp -size +100M 2>/dev/null

After freeing space, start the service:

sudo systemctl start mysql

Cause 2: Corrupt InnoDB redo logs

[ERROR] InnoDB: Log file ./ib_logfile0 is of different size
[ERROR] InnoDB: Plugin initialization aborted with error

Remove and regenerate the InnoDB log files (safe, they are regenerated automatically):

sudo systemctl stop mysql
sudo mv /var/lib/mysql/ib_logfile0 /tmp/ib_logfile0.bak
sudo mv /var/lib/mysql/ib_logfile1 /tmp/ib_logfile1.bak
sudo systemctl start mysql

If MySQL starts successfully, delete the backups:

rm /tmp/ib_logfile*.bak

Cause 3: Port already in use

[ERROR] Can't start server: Bind on TCP/IP port. Got error: 98: Address already in use

Find what's using port 3306:

sudo ss -tlnp | grep 3306

If another process is using it:

# Find and kill
sudo fuser -k 3306/tcp
sudo systemctl start mysql

If MySQL is already running as a zombie process:

sudo pkill mysqld
sudo systemctl start mysql

Cause 4: Wrong file permissions

[ERROR] Can't open and lock privilege tables: Table 'mysql.user' doesn't exist
[ERROR] mysqld: Can't create/write to file '/var/run/mysqld/mysqld.pid'

Fix ownership:

sudo chown -R mysql:mysql /var/lib/mysql
sudo chown -R mysql:mysql /var/run/mysqld
sudo chmod 755 /var/run/mysqld
sudo systemctl start mysql

If /var/run/mysqld doesn't exist:

sudo mkdir -p /var/run/mysqld
sudo chown mysql:mysql /var/run/mysqld
sudo chmod 755 /var/run/mysqld
sudo systemctl start mysql

Cause 5: Corrupted system tables

[ERROR] Fatal error: Can't open and lock privilege tables: Incorrect file format 'user'

Run the repair tool:

sudo systemctl stop mysql

# MySQL
sudo mysqld --tc-heuristic-recover=ROLLBACK

# MariaDB
sudo mysqld --tc-heuristic-recover=ROLLBACK

sudo systemctl start mysql

If that fails, repair with --skip-grant-tables:

sudo mysqld_safe --skip-grant-tables &
sleep 5
mysql -u root -e "REPAIR TABLE mysql.user;"
mysql -u root -e "FLUSH PRIVILEGES;"
sudo killall mysqld
sudo systemctl start mysql

Cause 6: Configuration syntax error

[ERROR] unknown variable 'innodb_buffer_pool_size1=128M'

Test the config file for syntax errors:

# MySQL
mysqld --validate-config

# MariaDB
mysqld --help --verbose 2>&1 | head -20

Then check /etc/mysql/mysql.conf.d/mysqld.cnf or /etc/mysql/mariadb.conf.d/50-server.cnf for typos.

Cause 7: AppArmor blocking MySQL

[ERROR] Could not open file '/var/lib/mysql-files/' for error logging: Permission denied

Check AppArmor:

sudo aa-status | grep mysql
sudo journalctl -k | grep DENIED | grep mysql

If AppArmor is blocking it:

# Temporarily disable MySQL AppArmor profile
sudo aa-disable /etc/apparmor.d/usr.sbin.mysqld
sudo systemctl start mysql

For a proper fix, update the AppArmor profile instead of disabling it.

Emergency: force InnoDB recovery

If none of the above works and you suspect severe InnoDB corruption, add to /etc/mysql/mysql.conf.d/mysqld.cnf:

[mysqld]
innodb_force_recovery = 1

Try starting with levels 1 through 6 until it starts. Once started, immediately dump all databases:

mysqldump --all-databases > /root/full-backup-$(date +%F).sql

Then remove the recovery option and restart cleanly.

innodb_force_recovery mode is read-only. Do not use it for normal operation. Increase the level only if the previous level failed to start the service.

Quick reference

Error keyword	Likely cause	Fix
`No space left`	Disk full	Free disk space
`ib_logfile`	Corrupt InnoDB logs	Remove ib_logfile*
`Address already in use`	Port conflict	Kill other process
`Permission denied`	Wrong ownership	`chown -R mysql:mysql`
`unknown variable`	Config typo	Fix `mysqld.cnf`
`Can't open privilege tables`	Missing data dir	Reinitialize or restore

On this page