Documentation Home
MySQL 8.0 Reference Manual
Related Documentation Download this Manual
PDF (US Ltr) - 37.8Mb
PDF (A4) - 37.9Mb
PDF (RPM) - 32.9Mb
HTML Download (TGZ) - 8.0Mb
HTML Download (Zip) - 8.1Mb
HTML Download (RPM) - 6.9Mb
Man Pages (TGZ) - 132.9Kb
Man Pages (Zip) - 189.5Kb
Info (Gzip) - 3.4Mb
Info (Zip) - 3.4Mb
Excerpts from this Manual

MySQL 8.0 Reference Manual  /  ...  /  Optimizing InnoDB Redo Logging

8.5.4 Optimizing InnoDB Redo Logging

Consider the following guidelines for optimizing redo logging:

  • Make your redo log files big, even as big as the buffer pool. When InnoDB has written the redo log files full, it must write the modified contents of the buffer pool to disk in a checkpoint. Small redo log files cause many unnecessary disk writes. Although historically big redo log files caused lengthy recovery times, recovery is now much faster and you can confidently use large redo log files.

    The size and number of redo log files are configured using the innodb_log_file_size and innodb_log_files_in_group configuration options. For information about modifying an existing redo log file configuration, see Section 15.7.2, “Changing the Number or Size of InnoDB Redo Log Files”.

  • Consider increasing the size of the log buffer. A large log buffer enables large transactions to run without a need to write the log to disk before the transactions commit. Thus, if you have transactions that update, insert, or delete many rows, making the log buffer larger saves disk I/O. Log buffer size is configured using the innodb_log_buffer_size configuration option, which can be configured dynamically in MySQL 8.0.

  • Configure the innodb_log_write_ahead_size configuration option to avoid read-on-write. This option defines the write-ahead block size for the redo log. Set innodb_log_write_ahead_size to match the operating system or file system cache block size. Read-on-write occurs when redo log blocks are not entirely cached to the operating system or file system due to a mismatch between write-ahead block size for the redo log and operating system or file system cache block size.

    Valid values for innodb_log_write_ahead_size are multiples of the InnoDB log file block size (2n). The minimum value is the InnoDB log file block size (512). Write-ahead does not occur when the minimum value is specified. The maximum value is equal to the innodb_page_size value. If you specify a value for innodb_log_write_ahead_size that is larger than the innodb_page_size value, the innodb_log_write_ahead_size setting is truncated to the innodb_page_size value.

    Setting the innodb_log_write_ahead_size value too low in relation to the operating system or file system cache block size results in read-on-write. Setting the value too high may have a slight impact on fsync performance for log file writes due to several blocks being written at once.

  • Optimize the use of spin delay by user threads waiting for flushed redo. Spin delay helps reduce latency. During periods of low concurrency, reducing latency may be less of a priority, and avoiding the use of spin delay during these periods may reduce energy consumption. During periods of high concurrency, you may want to avoid expending processing power on spin delay so that it can be used for other work. The following system variables permit setting high and low watermark values that define boundaries for the use of spin delay.

    • innodb_log_wait_for_flush_spin_hwm: Defines the maximum average log flush time beyond which user threads no longer spin while waiting for flushed redo. The default value is 400 microseconds.

    • innodb_log_spin_cpu_abs_lwm: Defines the minimum amount of CPU usage below which user threads no longer spin while waiting for flushed redo. The value is expressed as a sum of CPU core usage. For example, The default value of 80 is 80% of a single CPU core. On a system with a multi-core processor, a value of 150 represents 100% usage of one CPU core plus 50% usage of a second CPU core.

    • innodb_log_spin_cpu_pct_hwm: Defines the maximum amount of CPU usage above which user threads no longer spin while waiting for flushed redo. The value is expressed as a percentage of the combined total processing power of all CPU cores. The default value is 50%. For example, 100% usage of two CPU cores is 50% of the combined CPU processing power on a server with four CPU cores.

      The innodb_log_spin_cpu_pct_hwm configuration option respects processor affinity. For example, if a server has 48 cores but the mysqld process is pinned to only four CPU cores, the other 44 CPU cores are ignored.


User Comments
Sign Up Login You must be logged in to post a comment.