The index servers in a distributed system have one of the following roles:
A master index server is responsible for indexing. In the default configuration, it is not responsible for searching.
A slave index server is responsible only for searching and not for indexing.
The separation of the master index server and slave index servers is beneficial to performance. The indexing functions are separate from the searching functions, so that there is no loss of performance during indexing runs.
A backup index server can replace a master index server if it becomes unavailable. The backup index server is inactive if the master index server is available. When the master index server restarts after becoming unavailable, it takes over its tasks from the backup server again.
You implement backup index servers in order to make indexing highly available. The indexes must be stored centrally, so that both the master and the backup index servers can have write-access to them.
The index servers are the central components of a TREX system. In principle, their role determines the load that a host has to carry. The documentation below therefore refers to the hosts according to the role of the index server: A master, slave, or backup host is a host on which a master, slave, or backup index server is running.