12.6 Maintenance of a High Availability DB System
Oracle maintains a high availability DB System using a rolling maintenance process that has minimal downtime. The maintenance starts within two hours of the Maintenance window start time that you defined for the DB system.
The rolling maintenance process involves the following steps and does not require any input from you:
- One of the secondary nodes is removed from the replication group.
-
The maintenance process is performed on the selected node, which is returned to the replication group when the maintenance completes.
The returned secondary is allowed to catch up with the transactions that occurred while it was being maintained, before the maintenance of the remaining secondary begins.
- The remaining secondary node is removed from the replication group.
-
The maintenance process is performed on the selected node, which is returned to the replication group when the maintenance completes.
The returned secondary is allowed to catch up with the transactions that occurred while it was being maintained, before the next step begins.
- Existing connections to the primary are closed and no new connections
are permitted. The primary node is removed from the group and one of the secondary
nodes that have gone through maintenance is promoted to be the primary, like what
happens during a Failover. It is during this step that a brief period of downtime occurs
before the newly promoted primary resumes connections.
Note:
The current primary placement changes to that of the new primary. The preferred primary placement remains unchanged.Note:
The current binary log file name and position of the new primary may be different from the old primary's. 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. - The maintenance process is performed on the old primary node, which is returned to the replication group as a secondary node. No switching back to the old primary node is performed, to avoid additional downtime.
Parent topic: High Availability