Month: January 2026

MariaDB has many options for both the client and server than can be adjusted (or tuned) to provide better performance. For most, the default settings are fine and never need to be adjusted. A few, however, should be adjusted for optimal performance. This is particular true for the InnoDB storage engine. The option innodb_buffer_pool_size, which sets the size of the in-memory buffer pool, is probably the most important option to boost performance. It can be set up to 80% of the total memory on the server, but the default is very inadequate 128 MB.

Another option is max_allowed_packet which is used by both client and the server. This sets the maximum size for a data packet or generated/intermediate string. It must be large enough to hold the largest record or BLOB. The client default is not too shabby at 1 GB, but the server default is an anemic 16 MB.

The MariaDB packages for Debian and Red Hat based distributions do install options files (also known as defaults files) for both the client and server, but they are largely placeholders, setting very few or no options. And they have no comments about any of the options. This makes it difficult to know which options should be adjusted for best performance. The options are documented on the MariaDB Documentation site, but this means wading through a long list of options trying to find the ones of interest.

Read More

Having regular backups of MySQL and MariaDB databases is vitally important to ensure that data can be recovered in the event of corruption (which could be caused by a system failure or malicious software such as ransomware) or human error (such as when a user accidentally deletes a row, table, or entire database). MySQL and MariaDB provide a utility program to back up the databases – namely mysqldump (which is mariadb-dump for MariaDB 1.5 and higher) – but neither provides a program or script that allows backups to be easily automated. Aptum Technologies (my employer) provides a script allowing customers to back up the databases from a cron job. The original was a very basic script which didn’t have much flexibility. Using this as foundation, I rewrote the backup script and enhanced it over the years to improve its functionality and add features.

The script – backup-mysql – uses mysqldump to create a logical backup of each database. Each database is backed up to a separate file and then compressed using gzip (or pigz on systems with two or more processors). The database backup files are then archived to a tar file. This archive file is not compressed since the individual database backup files are already compressed. And a log file is created with information about the backup operation. Any errors that occur are recorded in the log file.

Read More