Reduce the Impact of Configuration Changes
As a distributed database, MarkLogic has to have a reliable system to keep all hosts in a cluster up to date when configuration changes happen. MarkLogic also has many configuration options. From group-level cache settings to app server threads to databases and what indexes are configured, all configuration has to be tracked and managed.
In versions prior to MarkLogic 11.2, when any configuration change happens, each host in the cluster reloads the configuration, clears many internal caches and causes a process called “index detection” to occur. When done in production, this process can have significant impact on performance if the system is under heavy load. This happens even if when configuration changes happen that should not require index detection to occur. MarkLogic 11.2 lays the groundwork for more granular handling of configuration changes to prevent this from happening.
In MarkLogic 11.2, when scheduled backups are configured for a database, MarkLogic will recognize that this is not a configuration change that requires clearing of caches and index detection. Therefore, systems that are running in production can modify the scheduled backup configuration without impacting the performance of user queries. With this framework in place, more granular handling of other configuration changes that should not require index detection will be added in future releases.