You can manually promote one of the secondary MySQL instances as
the primary instance. This is called as switchover.
When you perform a switchover, the preferred placement changes to the newly
selected placement of the primary instance, that is, the current placement of the
primary instance is same as the preferred placement. The IP address of the DB system
endpoint does not change. Switchover results in a short period of downtime while the
primary instance is redirected to the newly promoted instance. You must reopen the
database connections.
Following is the switchover process of a DB system that has been created or upgraded after February 25, 2025:
New transactions are blocked from the time the switchover is initiated.
Running transactions are allowed to continue to execute. After 5 minutes, any running transactions that have not completed are aborted and rolled back.
Immediately after all running transactions have completed or aborted, the DB system endpoint is assigned to the newly promoted primary instance.
Existing connections to the previous primary instance are closed. The client application must reopen new connections.
Following is the switchover process of a DB system that has been created before February 25, 2025 and has not been upgraded after February 25, 2025:
Running transactions are allowed to complete. The switchover process waits until all running transactions finish and commit.
New transactions are accepted and can perform reads and writes but, when the MySQL instance is promoted to be the new primary instance, and the connection to the previous primary instance is broken, all uncommitted transactions are rolled back.
The DB system endpoint is assigned to the newly promoted primary instance.
Existing connections to the previous primary instance are closed. The client application must reopen new connections.
Note
After a switchover, the current binary log file name and position of the new primary may be different from the old primary. As the binary logs of each instance are managed independently, each transaction recorded in the binary logs may be written to a different binary log file and position in different instances.
Switchover Scenarios
You can use switchover in the following scenarios:
Application testing: Ensuring your application works correctly
with the newly promoted primary instance.
Proximity of availability domain: While fault domains are
contained in the same data center, availability domains are spread across different
data centers. Even though these availability domains are connected with a
low-latency network, it may be necessary to test that latency by moving primary
instance from availability domain to availability domain to confirm the best
location for your primary instance. For example, switching to an instance in the
same availability domain as the application which connects to its.
HeatWave Cluster Support
When you perform a switchover, the HeatWave cluster attached to the earlier
primary instance is deleted and a new HeatWave cluster is attached to the new primary
instance.
When you perform a switchover, a long running query offloaded to the HeatWave cluster can get aborted. In this case, you need to reconnect and rerun the query after the HeatWave cluster is ready.
Do the following to switch from the current primary instance to one of the
secondary instances:
Open the command-line interface and list all the availability domains in the
current region:
oci iam availability-domain list
Switchover to the desired MySQL instance:
For multi-availability domain:
oci mysql db-system update --db-system-id <DBSystemOCID> --availability-domain <SelectedAvailabilityDomain>
For single-availability
domain:
oci mysql db-system update --db-system-id <DBSystemOCID> --fault-domain <SelectedFaultDomain>
Where
db-system-id: Specify the OCID of the DB
system.
availability-domain: Specify the name of
the availability domain in which the secondary MySQL instance, which you
want to promote to the primary, is present.
fault-domain: Specify the name of the
fault domain in which the secondary MySQL instance, which you want to
promote to the primary, is present.
The status of the DB system changes to Updating, and the instance in the
selected availability or fault domain becomes the primary instance and the read/write
endpoint.