=========================== Set up database replication =========================== You can create a replica of an existing database instance. When you make subsequent changes to the original instance, the system automatically applies those changes to the replica. - Replicas are read-only. - When you create a replica, do not specify the ``--users`` or ``--databases`` options. - You can choose a smaller volume or flavor for a replica than for the original, but the replica's volume must be big enough to hold the data snapshot from the original. This example shows you how to replicate a MySQL database instance. Set up replication ~~~~~~~~~~~~~~~~~~ #. **Get the instance ID** Get the ID of the original instance you want to replicate: .. code:: $ trove list +-----------+------------+-----------+-------------------+--------+-----------+------+ | id | name | datastore | datastore_version | status | flavor_id | size | +-----------+------------+-----------+-------------------+--------+-----------+------+ | 97b...ae6 | base_1 | mysql | mysql-5.5 | ACTIVE | 10 | 2 | +-----------+------------+-----------+-------------------+--------+-----------+------+ #. **Create the replica** Create a new instance that will be a replica of the original instance. You do this by passing in the ``--replica_of`` option with the trove ``create`` command. This example creates a replica called ``replica_1``. ``replica_1`` is a replica of the original instance, ``base_1``: .. code:: $ trove create replica_1 6 --size=5 --datastore_version mysql-5.5 \ --datastore mysql --replica_of ID_OF_ORIGINAL_INSTANCE #. **Verify replication status** Pass in ``replica_1``'s instance ID with the trove ``show`` command to verify that the newly created ``replica_1`` instance is a replica of the original ``base_1``. Note that the ``replica_of`` property is set to the ID of ``base_1``. .. code:: $ trove show INSTANCE_ID_OF_REPLICA_1 +-------------------+--------------------------------------+ | Property | Value | +-------------------+--------------------------------------+ | created | 2014-09-16T11:16:49 | | datastore | mysql | | datastore_version | mysql-5.5 | | flavor | 6 | | id | 49c6eff6-ef91-4eff-91c0-efbda7e83c38 | | name | replica_1 | | replica_of | 97b4b853-80f6-414f-ba6f-c6f455a79ae6 | | status | BUILD | | updated | 2014-09-16T11:16:49 | | volume | 5 | +-------------------+--------------------------------------+ Now pass in ``base_1``'s instance ID with the trove ``show`` command to list the replica(s) associated with the original instance. Note that the ``replicas`` property is set to the ID of ``replica_1``. If there are multiple replicas, they appear as a comma-separated list. .. code:: $ trove show INSTANCE_ID_OF_BASE_1 +-------------------+--------------------------------------+ | Property | Value | +-------------------+--------------------------------------+ | created | 2014-09-16T11:04:56 | | datastore | mysql | | datastore_version | mysql-5.5 | | flavor | 6 | | id | 97b4b853-80f6-414f-ba6f-c6f455a79ae6 | | ip | 172.16.200.2 | | name | base_1 | | replicas | 49c6eff6-ef91-4eff-91c0-efbda7e83c38 | | status | ACTIVE | | updated | 2014-09-16T11:05:06 | | volume | 5 | | volume_used | 0.11 | +-------------------+--------------------------------------+ #. **Detach the replica** If the original instance goes down, you can detach the replica. The replica becomes a standalone database instance. You can then take the new standalone instance and create a new replica of that instance. You detach a replica using the trove ``detach-replica`` command: .. code:: $ trove detach-replica INSTANCE_ID_OF_REPLICA