GLPI can be configured to use replica databases in some cases to improve performance.
This tab is only visible if the option has been enabled in the system tab.
To activate the replica management, you just have to fill in the connection parameters. It is advisable to use a login that has only read privileges on the database.
You can use several replicates by separating them with a blank space. As the result in the config_db_slave.php file, the dbhost parameter will be an array.
A specific port can also be added.
If you use multiple replicas, the settings will apply to them all.
The SQL replicas will be used in at least the following cases:
- The main server cannot be reached. In this case, GLPI will switch to read-only mode.
- Running reports to reduce the load on the main server.
It is also possible to use replicas for the search engine depending on the option chosen.
- If synced (all changes) GLPI will check the synchronization and use the replica if all changes have been replicated.
- If synced (current user changes) GLPI will check the synchronization and use the replica if all the changes for the current user have been replicated. If using a read-only account, the replica will always be used.
- If synced or read-only account GLPI will check the synchronization and use the replica if all changes have been replicated. If using a read-only account, the replica will always be used.
When always using the replica, it is possible that the results will be out of date. A different option should be considered before this one.
The replication configuration of your database is not set by GLPI. You still need to configure that with the appropriate tools for your database management system.