RDS DB Instance Maintenance & Upgrades
A DB instance may be subject to changes if it is modified manually, for example. Amazon RDS performs maintenance or upgrades to the DB engine.
Amazon RDS Maintenance
Amazon RDS periodically performs maintenance on Amazon RDS resource, such as DB instances. This usually involves updates to the DB instance’s operating system (OS).
You can apply maintenance items manually to a DB instance at your convenience.
Or, wait for Amazon RDS to initiate the automatic maintenance process during the defined weekly maintenance window.
The maintenance window determines when operations are started, but it does not limit the execution time. Maintenance operations can continue beyond the end of the maintenance window, but they are not guaranteed to be completed before that time.
You can check the RDS console for updates and the RDS API to verify that they are available. If an update is available one can access the maintenance items.
Apply the maintenance items immediately.
They should be scheduled to begin during the next defined maintenance window
Maintenance items marked asRequired can’t be deferred indefinitely. If deferred AWS will send an email letting you know when the next update will be performed.
Available and can be deferred indefinitely. The update will not apply to the DB instance.
Only patches that affect security and reliability are automatically scheduled for patching. This patching is rare (typically once a few months) and rarely requires more than a fraction your maintenance window.
RDS will temporarily take your DB instance offline in order to perform maintenance items. Scale compute operations and database patching are two examples of maintenance that require DB instances to be offline. These operations typically take a few minutes.
Multi-AZ deployment of the DB instance reduces the impact on a maintenance event. Follow these steps: Perform maintenance on the standby.
Promote the standby to primary.
Perform maintenance on the primary and then switch to the standby.
RDS upgrades both primary and secondary DB instances simultaneously when the database engine for the DB instance in a MultiAZ deployment is modified. The entire Multi-AZ deployment will be shut down during an upgrade.
Most often, upgrades to the operating system are for security reasons and should be completed as soon as possible.
OS updates for a DB instance are available at one’s convenience or you can wait for RDS to initiate the maintenance process to apply the update during the specified maintenance window
DB instance is not automatically backed-up when an OS update occurs. Backups should be done before the update is applied.
When RDS supports a new DB engine, the DB instance engine version can also be upgraded.
Major and minor database version upgrades can contain changes that aren’t backward-compatible
RDS does not automatically apply major version updates
Before deploying it to production, the DB instance must be manually modified.
Minor version upgradesEach DB Engine handles minor version upgrades differently, for e.g. RDS applies minor version upgrades automatically to a DB server running PostgreSQL. However, it must be manually applied for a DB server running Oracle.
Amazon posts an announcement on the forums announcement page and sends a customer an e-mail notification prior to upgrading a DB instance.
Amazon schedules the upgrades at certain times throughout the year to help you plan around them. Downtime is required to upgrade a DB Engine version, even for MultiAZ instances.
RDS takes two DB snaps during an upgrade. The first is of the DB instance.