Maintaining an Amazon Aurora DB cluster

Periodically, Amazon RDS performs maintenance on Amazon RDS resources. The following topics describe these maintenance actions and how to apply them.

Overview of DB cluster maintenance updates

Maintenance most often involves updates to the following resources in your DB cluster:

Updates to the operating system most often occur for security issues. We recommend that you do them as soon as possible. For more information about operating system updates, see Operating system updates for Aurora DB clusters.

Offline resources during maintenance updates

Some maintenance items require that Amazon RDS take your DB cluster offline for a short time. Maintenance items that require a resource to be offline include required operating system or database patching. Required patching is automatically scheduled only for patches that are related to security and instance reliability. Such patching occurs infrequently, typically once every few months. It seldom requires more than a fraction of your maintenance window.

Deferred DB instance and DB cluster modifications

Deferred DB cluster and instance modifications that you have chosen not to apply immediately are applied during the maintenance window. For example, you might choose to change DB instance classes or cluster or DB parameter groups during the maintenance window. Such modifications that you specify using the pending reboot setting don't show up in the Pending maintenance list. For information about modifying a DB cluster, see Modifying an Amazon Aurora DB cluster.

To see the modifications that are pending for the next maintenance window, use the describe-db-clusters AWS CLI command and check the PendingModifiedValues field.

Eventual consistency for the DescribePendingMaintenanceActions API

The Amazon RDS DescribePendingMaintenanceActions API follows an eventual consistency model. This means that the result of the DescribePendingMaintenanceActions command might not be immediately visible to all subsequent RDS commands. Keep this in mind when you use DescribePendingMaintenanceActions immediately after using a previous API command.

Eventual consistency can affect the way you managed your maintenance updates. For example, if you run the ApplyPendingMaintenanceActions command to update the database engine version for a DB cluster, it will eventually be visible to DescribePendingMaintenanceActions. In this scenario, DescribePendingMaintenanceActions might show that the maintenance action wasn't applied even though it was.

To manage eventual consistency, you can do the following:

Viewing pending maintenance updates

View whether a maintenance update is available for your DB cluster by using the RDS console, the AWS CLI, or the RDS API. If an update is available, it is indicated in the Maintenance column for the DB cluster on the Amazon RDS console, as shown in this figure.

If no maintenance update is available for a DB cluster, the column value is none for it.

If a maintenance update is available for a DB cluster, the following column values are possible:

If an update is available, you can do one of the following:

The maintenance window determines when pending operations start, but doesn't limit the total run time of these operations. Maintenance operations aren't guaranteed to finish before the maintenance window ends, and can continue beyond the specified end time. For more information, see Amazon RDS maintenance window.

You can also view whether a maintenance update is available for your DB cluster by running the describe-pending-maintenance-actions AWS CLI command.

For information about applying maintenance updates, see Applying updates to a DB cluster.

Maintenance actions for Amazon Aurora

The following maintenance actions apply to Aurora DB clusters:

The following maintenance actions apply to Aurora DB instances:

Choosing the frequency of Aurora MySQL maintenance updates

You can control whether Aurora MySQL upgrades happen frequently or rarely for each DB cluster. The best choice depends on your usage of Aurora MySQL and the priorities for your applications that run on Aurora. For information about the Aurora MySQL long-term stability (LTS) releases that require less frequent upgrades, see Aurora MySQL long-term support (LTS) releases.

You might choose to upgrade an Aurora MySQL cluster rarely if some or all of the following conditions apply:

If the preceding factors apply to your situation, you can limit the number of forced upgrades for an Aurora MySQL DB cluster. You do so by choosing a specific Aurora MySQL version known as the "Long-Term Support" (LTS) version when you create or upgrade that DB cluster. Doing so minimizes the number of upgrade cycles, testing cycles, and upgrade-related outages for that DB cluster.

You might choose to upgrade an Aurora MySQL cluster frequently if some or all of the following conditions apply:

If the preceding factors apply to your situation, you can enable Aurora to apply important upgrades more frequently. To do so, upgrade an Aurora MySQL DB cluster to a more recent Aurora MySQL version than the LTS version. Doing so makes the latest performance enhancements, bug fixes, and features available to you more quickly.

Amazon RDS maintenance window

The maintenance window is a weekly time interval during which any system changes are applied. Every DB cluster has a weekly maintenance window. The maintenance window is an opportunity to control when modifications and software patching occur. For more information about adjusting the maintenance window, see Adjusting the preferred DB cluster maintenance window.

RDS consumes some of the resources on your DB cluster while maintenance is being applied. You might observe a minimal effect on performance. For a DB instance, on rare occasions, a Multi-AZ failover might be required for a maintenance update to complete.

If a maintenance event is scheduled for a given week, it's initiated during the 30-minute maintenance window you identify. Most maintenance events also complete during the 30-minute maintenance window, although larger maintenance events may take more than 30 minutes to complete. The maintenance window is paused when the DB cluster is stopped.

The 30-minute maintenance window is selected at random from an 8-hour block of time per region. If you don't specify a maintenance window when you create the DB cluster, RDS assigns a 30-minute maintenance window on a randomly selected day of the week.

The following table shows the time blocks for each AWS Region from which default maintenance windows are assigned.

Adjusting the preferred DB cluster maintenance window

The Aurora DB cluster maintenance window should fall at the time of lowest usage and thus might need modification from time to time. Your DB cluster is unavailable during this time only if the updates that are being applied require an outage. The outage is for the minimum amount of time required to make the necessary updates.

aws rds modify-db-cluster \
--db-cluster-identifier my-cluster \
--preferred-maintenance-window Tue:04:00-Tue:04:30

aws rds modify-db-cluster ^
--db-cluster-identifier my-cluster ^
--preferred-maintenance-window Tue:04:00-Tue:04:30

Applying updates to a DB cluster

With Amazon RDS, you can choose when to apply maintenance operations. You can decide when Amazon RDS applies updates by using the AWS Management Console, AWS CLI, or RDS API.

aws rds apply-pending-maintenance-action \
    --resource-identifier arn:aws:rds:us-west-2:001234567890:db:mysql-db \
    --apply-action system-update \
    --opt-in-type immediate

aws rds apply-pending-maintenance-action ^
    --resource-identifier arn:aws:rds:us-west-2:001234567890:db:mysql-db ^
    --apply-action system-update ^
    --opt-in-type immediate

aws rds describe-pending-maintenance-actions \
    --resource-identifier arn:aws:rds:us-west-2:001234567890:db:mysql-db

aws rds describe-pending-maintenance-actions ^
    --resource-identifier arn:aws:rds:us-west-2:001234567890:db:mysql-db

aws rds describe-pending-maintenance-actions \
	--filters Name=db-cluster-id,Values=sample-cluster1,sample-cluster2

aws rds describe-pending-maintenance-actions ^
	--filters Name=db-cluster-id,Values=sample-cluster1,sample-cluster2

Automatic minor version upgrades for Aurora DB clusters

The Auto minor version upgrade setting specifies whether Aurora automatically applies upgrades to your DB cluster. These upgrades include new minor versions containing additional features and patches containing bug fixes.

This setting is turned on by default. For each new DB cluster, choose the appropriate value for this setting. This value is based on its importance, expected lifetime, and the amount of verification testing that you do after each upgrade.

For instructions on turning the Auto minor version upgrade setting on or off, see the following:

The following table shows how the Auto minor version upgrade setting works when applied at the cluster and instance levels.

Automatic minor version upgrades are communicated in advance through an Amazon RDS DB cluster event with a category of maintenance and ID of RDS-EVENT-0156. For more information, see Amazon RDS event categories and event messages for Aurora.

Automatic upgrades occur during the maintenance window. If the individual DB instances in the DB cluster have different maintenance windows from the cluster maintenance window, then the cluster maintenance window takes precedence.

For more information about engine updates for Aurora PostgreSQL, see Database engine updates for Amazon Aurora PostgreSQL.

For more information about the Auto minor version upgrade setting for Aurora MySQL, see Enabling automatic upgrades between minor Aurora MySQL versions. For general information about engine updates for Aurora MySQL, see Database engine updates for Amazon Aurora MySQL.

You can use a CLI command such as the following to check the status of the AutoMinorVersionUpgrade setting for all of the DB instances in your Aurora MySQL clusters.

aws rds describe-db-instances \
  --query '*[].{DBClusterIdentifier:DBClusterIdentifier,DBInstanceIdentifier:DBInstanceIdentifier,AutoMinorVersionUpgrade:AutoMinorVersionUpgrade}'

That command produces output similar to the following:

[
  {
      "DBInstanceIdentifier": "db-writer-instance",
      "DBClusterIdentifier": "my-db-cluster-57",
      "AutoMinorVersionUpgrade": true
  },
  {
      "DBInstanceIdentifier": "db-reader-instance1",
      "DBClusterIdentifier": "my-db-cluster-57",
      "AutoMinorVersionUpgrade": false
  },
  {
      "DBInstanceIdentifier": "db-writer-instance2",
      "DBClusterIdentifier": "my-db-cluster-80",
      "AutoMinorVersionUpgrade": true
  },
... output omitted ...

In this example, Enable auto minor version upgrade is turned off for the DB cluster my-db-cluster-57, because it's turned off for one of the DB instances in the cluster.

Operating system updates for Aurora DB clusters

DB instances in Aurora MySQL and Aurora PostgreSQL DB clusters occasionally require operating system updates. Amazon RDS upgrades the operating system to a newer version to improve database performance and customers’ overall security posture. Typically, the updates take about 10 minutes. Operating system updates don't change the DB engine version or DB instance class of a DB instance.

There are two types of operating system updates, differentiated by the description for the pending maintenance action:

Operating system updates can be either optional or mandatory:

For Aurora DB clusters, you can perform updates at both the cluster and instance levels:

When you schedule operating system updates at the cluster level, the updates occur during the cluster-level maintenance window specified for the cluster. This ensures that updates are coordinated for the entire cluster. However, if you schedule updates at the instance level, the updates occur during the maintenance window of each respective DB instance. If he cluster and two reader instances have different maintenance window times, scheduling an operating system update at the cluster level aligns with the cluster maintenance window.

You can use the AWS Management Console or the AWS CLI to get information about the type of operating system upgrade.

To get update information using the AWS Management Console

1. Sign in to the AWS Management Console and open the Amazon RDS console at https://console.aws.amazon.com/rds/.

2. In the navigation pane, choose Databases, and then select the DB instance.

3. Choose Maintenance & backups.

4. In the Pending maintenance section, find the operating system update, and check the Description value.

The following images show a DB cluster with a writer DB instance that has an operating system patch available.

The following images show a DB cluster with a writer DB instance and a reader DB instance. The writer instance has a mandatory operating system upgrade available. The reader instance has an operating system patch available.

To get update information from the AWS CLI, use the describe-pending-maintenance-actions command.

aws rds describe-pending-maintenance-actions

The following output shows an operating system distribution upgrade for a DB cluster and a DB instance.

{
  "PendingMaintenanceActions": [
    {
      "ResourceIdentifier": "arn:aws:rds:us-east-1:123456789012:cluster:t3",
      "PendingMaintenanceActionDetails": [
        {
          "Action": "os-upgrade",
          "Description": "New Operating System upgrade is available"
        }
      ]
    },
    {
      "ResourceIdentifier": "arn:aws:rds:us-east-1:123456789012:db:t3-instance1",
      "PendingMaintenanceActionDetails": [
        {
          "Action": "system-update",
          "Description": "New Operating System upgrade is available"
        }
      ]
    },
  ]
}

The following output shows an operating system patch for a DB instance.

{
  "ResourceIdentifier": "arn:aws:rds:us-east-1:123456789012:db:mydb2",
  "PendingMaintenanceActionDetails": [
    {
      "Action": "system-update",
      "Description": "New Operating System patch is available"
    }
  ]
}

Availability of operating system updates

Operating system updates are specific to DB engine version and DB instance class. Therefore, DB instances receive or require updates at different times. When an operating system update is available for your DB instance based on its engine version and instance class, the update appears in the console. It can also be viewed by running the describe-pending-maintenance-actions AWS CLI command or by calling the DescribePendingMaintenanceActions RDS API operation. If an update is available for your instance, you can update your operating system by following the instructions in Applying updates to a DB cluster.