Making changes to a multi-database system that uses C Replicator (CREP) involves a number of tasks. The following sections summarize the tasks that you must perform for each specified scenario:
Adding a server to an existing replication network
To add a server to an existing replication network, complete the following tasks:
1 Install and start a new database and create a schema.
2 If the new database is going to be a source database, define a replication partition in the source database.
3 If the new database is going to be a target database, complete the following steps:
– Install and start a replicator instance.
– Configure the replicator for the target database. If required, perform initial load by using the LOAD ONLY option with the START REPLICATION statement.
4 If the database is a source database for other targets, execute appropriate replicator statements on all applicable replicator instances.
Removing a server from an existing replication network
To remove a server from an existing replication network, complete the following tasks:
1 Shut down the replicator for the target database.
2 Remove the database and delete the server.
3 To prevent source database(s) from storing transaction logs for a removed server, drop the partition.
Upgrading solidDB binaries in network nodes
It is possible to upgrade the solidDB binaries in each server or the replicator in the network without closing down the traffic in all nodes of the network. Complete the following process:
1 Stop all replicator activity related to the database instance.
2 Shut down the database instance, upgrade the solidDB binaries, and restart the database.
3 When the database has restarted, restart replication.
It is possible to upgrade a CREP-based network of servers one by one without affecting the availability of the entire network.
Note Upgrades that use this method are supported only within the same solidDB major release.
Changing the database schema
To change the database schema, complete the following steps:
1 Stop all replication activity in both directions.
2 Run the schema modification scripts on both source and target databases.
Note Some schema modifications (for example, CREATE INDEX) can take a long time.