Not a ground-breaking post here, but if you are interested in knowing more about the innodb_log_block_size variable, or if you use SSD cards and/or large InnoDB log files on ext4, then this is for you.

I’d read about it before briefly before, but didn’t give it too much thought until I ran across the following entry in an error log the other day:

InnoDB: Warning: innodb_log_block_size has been changed
from default value 512. (###EXPERIMENTAL### operation)

This got me wanting to know more.

Basically, this variable changes the size of transaction log records. Generally, the default of 512 is a good value. However, it has been found that setting it to 4096 has been beneficial when using SSD cards. (Note that while it is possible to set this to a value other than 512 or 4096, those are currently the only 2 values that make sense to use.)

Also, it has been found that 4096 is the best setting if you run ext4 along with innodb_flush_method=ALL_O_DIRECT (note both innodb_log_block_size and ALL_O_DIRECT are specific to XtraDB/XtraDB+; read: MariaDB and Percona server).

What is ALL_O_DIRECT and when it is useful?

“ALL_O_DIRECT: use O_DIRECT to open both data and log files, and use fsync() to flush the data files but not the log files. This option is recommended when InnoDB log files are big (more than 8GB), otherwise there might be even a performance degradation.”

http://www.percona.com/…/innodb_io.html#innodb_flush_method

Thus if you’re running large InnoDB log files (8G+) on ext4, you may want to consider ALL_O_DIRECT, and because you’re on ext4, you should set innodb_log_block_size=4096, the default log-block-size in ext4, in order to avoid the unaligned AIO/DIO warnings.

Likewise, if running with SSD cards, you’d likely see a performance improvement with innodb_log_block_size=4096 also.

If interested, there is a bit more about the innodb_log_block_size option here:

https://mariadb.com/…variables/#innodb_log_block_size
http://www.percona.com/…/innodb_io_55.html#innodb_log_block_size

Hope this helps.