Help Center/ TaurusDB/ Kernel/ Common Kernel Functions/ Fast Binlog Positioning
Updated on 2025-05-12 GMT+08:00
View PDF
Share

Fast Binlog Positioning

If you pull binlogs from a TaurusDB instance using auto-positioning, it can be time-consuming to locate the correct binlog position when there are a large number of unread binlogs. For example, if there are 60,000 unread binlogs, it would take about 60 minutes to automatically locate the correct binlog position. With fast binlog positioning enabled, the time required can be reduced to about 2 seconds.

Constraint

The kernel version must be 2.0.60.241200 or later.

Parameter Description

rds_locate_right_binlog_by_binary_search

  • Scope: Global
  • Dynamic: Yes
  • Type: Boolean
  • Default Value: true
  • Description: Controls whether the sending thread uses binary search to locate the correct binlog position when auto-positioning is used to set up replication.

Example

Create a TaurusDB instance (IP address: 192.168.187.120). Execute transactions on the TaurusDB instance to generate about 60,000 binlogs.

Create a DR instance and run the following commands on the DR instance to pull unread binlogs from the TaurusDB instance:

CHANGE MASTER TO MASTER_HOST='192.168.187.120', MASTER_PORT=3306, MASTER_AUTO_POSITION = 1;

START SLAVE user='rdsRepl' password='......';

SHOW SLAVE STATUS

Observe how long it takes to automatically locate the correct binlog position (from the time when START SLAVE is executed to the time when Retrieved_Gtid_Set starts to increase).

Retrieved_Gtid_Set is one of the items in the output of SHOW SLAVE STATUS.

Parent Topic: Common Kernel Functions