Using write forwarding in an Aurora MySQL global database

Region and version availability of write forwarding in Aurora MySQL

Write forwarding is supported with Aurora MySQL 2.08.1 and higher versions, in every Region where Aurora MySQL-based global databases are available.

For information on version and Region availability of Aurora MySQL global databases, see Aurora global databases with Aurora MySQL.

Enabling write forwarding in Aurora MySQL

By default, write forwarding isn't enabled when you add a secondary cluster to an Aurora global database.

To enable write forwarding using the Amazon Web Services Management Console, select the Turn on global write forwarding check box under Read replica write forwarding when you add a Region for a global database. For an existing secondary cluster, modify the cluster to Turn on global write forwarding. To turn off write forwarding, clear the Turn on global write forwarding check box when adding the Region or modifying the secondary cluster.

To enable write forwarding using the Amazon CLI, use the --enable-global-write-forwarding option. This option works when you create a new secondary cluster using the create-db-cluster command. It also works when you modify an existing secondary cluster using the modify-db-cluster command. It requires that the global database uses an Aurora version that supports write forwarding. You can turn write forwarding off by using the --no-enable-global-write-forwarding option with these same CLI commands.

To enable write forwarding using the Amazon RDS API, set the EnableGlobalWriteForwarding parameter to true. This parameter works when you create a new secondary cluster using the CreateDBCluster operation. It also works when you modify an existing secondary cluster using the ModifyDBCluster operation. It requires that the global database uses an Aurora version that supports write forwarding. You can turn write forwarding off by setting the EnableGlobalWriteForwarding parameter to false.

The following CLI examples show how you can set up an Aurora global database with write forwarding enabled or disabled. The highlighted items represent the commands and options that are important to specify and keep consistent when setting up the infrastructure for an Aurora global database.

The following example creates an Aurora global database, a primary cluster, and a secondary cluster with write forwarding enabled. Substitute your own choices for the user name, password, and primary and secondary Amazon Regions.

# Create overall global database.
aws rds create-global-cluster --global-cluster-identifier write-forwarding-test \
  --engine aurora-mysql --engine-version 5.7.mysql_aurora.2.11.1 \
  --region us-east-1

# Create primary cluster, in the same Amazon Region as the global database.
aws rds create-db-cluster --global-cluster-identifier write-forwarding-test \
  --db-cluster-identifier write-forwarding-test-cluster-1 \
  --engine aurora-mysql --engine-version 5.7.mysql_aurora.2.11.1 \
  --master-username user_name --master-user-password password \
  --region us-east-1

aws rds create-db-instance --db-cluster-identifier write-forwarding-test-cluster-1 \
  --db-instance-identifier write-forwarding-test-cluster-1-instance-1 \
  --db-instance-class db.r5.large \
  --engine aurora-mysql --engine-version 5.7.mysql_aurora.2.11.1 \
  --region us-east-1

aws rds create-db-instance --db-cluster-identifier write-forwarding-test-cluster-1 \
  --db-instance-identifier write-forwarding-test-cluster-1-instance-2 \
  --db-instance-class db.r5.large \
  --engine aurora-mysql --engine-version 5.7.mysql_aurora.2.11.1 \
  --region us-east-1

# Create secondary cluster, in a different Amazon Region than the global database,
# with write forwarding enabled.
aws rds create-db-cluster --global-cluster-identifier write-forwarding-test \
  --db-cluster-identifier write-forwarding-test-cluster-2 \
  --engine aurora-mysql --engine-version 5.7.mysql_aurora.2.11.1 \
  --region us-east-2 \
  --enable-global-write-forwarding

aws rds create-db-instance --db-cluster-identifier write-forwarding-test-cluster-2 \
  --db-instance-identifier write-forwarding-test-cluster-2-instance-1 \
  --db-instance-class db.r5.large \
  --engine aurora-mysql --engine-version 5.7.mysql_aurora.2.11.1 \
  --region us-east-2

aws rds create-db-instance --db-cluster-identifier write-forwarding-test-cluster-2 \
  --db-instance-identifier write-forwarding-test-cluster-2-instance-2 \
  --db-instance-class db.r5.large \
  --engine aurora-mysql --engine-version 5.7.mysql_aurora.2.11.1 \
  --region us-east-2

The following example continues from the previous one. It creates a secondary cluster without write forwarding enabled, then enables write forwarding. After this example finishes, all secondary clusters in the global database have write forwarding enabled.

# Create secondary cluster, in a different Amazon Region than the global database,
# without write forwarding enabled.
aws rds create-db-cluster --global-cluster-identifier write-forwarding-test \
  --db-cluster-identifier write-forwarding-test-cluster-2 \
  --engine aurora-mysql --engine-version 5.7.mysql_aurora.2.11.1 \
  --region us-west-1

aws rds create-db-instance --db-cluster-identifier write-forwarding-test-cluster-2 \
  --db-instance-identifier write-forwarding-test-cluster-2-instance-1 \
  --db-instance-class db.r5.large \
  --engine aurora-mysql --engine-version 5.7.mysql_aurora.2.11.1 \
  --region us-west-1

aws rds create-db-instance --db-cluster-identifier write-forwarding-test-cluster-2 \
  --db-instance-identifier write-forwarding-test-cluster-2-instance-2 \
  --db-instance-class db.r5.large \
  --engine aurora-mysql --engine-version 5.7.mysql_aurora.2.11.1 \
  --region us-west-1

aws rds modify-db-cluster --db-cluster-identifier write-forwarding-test-cluster-2 \
  --region us-east-2 \
  --enable-global-write-forwarding

Checking if a secondary cluster has write forwarding enabled in Aurora MySQL

To determine whether you can use write forwarding from a secondary cluster, you can check whether the cluster has the attribute "GlobalWriteForwardingStatus": "enabled".

In the Amazon Web Services Management Console, on the Configuration tab of the details page for the cluster, you see the status Enabled for Global read replica write forwarding.

To see the status of the global write forwarding setting for all of your clusters, run the following Amazon CLI command.

A secondary cluster shows the value "enabled" or "disabled" to indicate if write forwarding is turned on or off. A value of null indicates that write forwarding isn't available for that cluster. Either the cluster isn't part of a global database, or is the primary cluster instead of a secondary cluster. The value can also be "enabling" or "disabling" if write forwarding is in the process of being turned on or off.

aws rds describe-db-clusters \
--query '*[].{DBClusterIdentifier:DBClusterIdentifier,GlobalWriteForwardingStatus:GlobalWriteForwardingStatus}'

[
    {
        "GlobalWriteForwardingStatus": "enabled",
        "DBClusterIdentifier": "aurora-write-forwarding-test-replica-1"
    },
    {
        "GlobalWriteForwardingStatus": "disabled",
        "DBClusterIdentifier": "aurora-write-forwarding-test-replica-2"
    },
    {
        "GlobalWriteForwardingStatus": null,
        "DBClusterIdentifier": "non-global-cluster"
    }
]

To find all secondary clusters that have global write forwarding enabled, run the following command. This command also returns the cluster's reader endpoint. You use the secondary cluster's reader endpoint when you use write forwarding from the secondary to the primary in your Aurora global database.

aws rds describe-db-clusters --query 'DBClusters[].{DBClusterIdentifier:DBClusterIdentifier,GlobalWriteForwardingStatus:GlobalWriteForwardingStatus,ReaderEndpoint:ReaderEndpoint} | [?GlobalWriteForwardingStatus == `enabled`]'
[
    {
        "GlobalWriteForwardingStatus": "enabled",
        "ReaderEndpoint": "aurora-write-forwarding-test-replica-1.cluster-ro-cnpexample.us-west-2.rds.amazonaws.com",
        "DBClusterIdentifier": "aurora-write-forwarding-test-replica-1"
    }
]

Application and SQL compatibility with write forwarding in Aurora MySQL

You can use the following kinds of SQL statements with write forwarding:

Certain statements aren't allowed or can produce stale results when you use them in a global database with write forwarding. Thus, the EnableGlobalWriteForwarding setting is turned off by default for secondary clusters. Before turning it on, check to make sure that your application code isn't affected by any of these restrictions.

The following restrictions apply to the SQL statements you use with write forwarding. In some cases, you can use the statements on secondary clusters with write forwarding enabled at the cluster level. This approach works if write forwarding isn't turned on within the session by the aurora_replica_read_consistency configuration parameter. Trying to use a statement when it's not allowed because of write forwarding causes an error message with the following format.

ERROR 1235 (42000): This version of MySQL doesn't yet support 'operation with write forwarding'.

Isolation and consistency for write forwarding in Aurora MySQL

In sessions that use write forwarding, you can only use the REPEATABLE READ isolation level. Although you can also use the READ COMMITTED isolation level with read-only clusters in secondary Amazon Regions, that isolation level doesn't work with write forwarding. For information about the REPEATABLE READ and READ COMMITTED isolation levels, see Aurora MySQL isolation levels.

You can control the degree of read consistency on a secondary cluster. The read consistency level determines how much waiting the secondary cluster does before each read operation to ensure that some or all changes are replicated from the primary cluster. You can adjust the read consistency level to ensure that all forwarded write operations from your session are visible in the secondary cluster before any subsequent queries. You can also use this setting to ensure that queries on the secondary cluster always see the most current updates from the primary cluster. This is so even for those submitted by other sessions or other clusters. To specify this type of behavior for your application, you choose a value for the session-level parameter aurora_replica_read_consistency.

For the aurora_replica_read_consistency parameter, you can specify the values EVENTUAL, SESSION, and GLOBAL.

As you increase the consistency level, your application spends more time waiting for changes to be propagated between Amazon Regions. You can choose the balance between fast response time and ensuring that changes made in other locations are fully available before your queries run.

With the read consistency set to EVENTUAL, queries in a secondary Amazon Region that uses write forwarding might see data that is slightly stale due to replication lag. Results of write operations in the same session aren't visible until the write operation is performed on the primary Region and replicated to the current Region. The query doesn't wait for the updated results to be available. Thus, it might retrieve the older data or the updated data, depending on the timing of the statements and the amount of replication lag.

With the read consistency set to SESSION, all queries in a secondary Amazon Region that uses write forwarding see the results of all changes made in that session. The changes are visible regardless of whether the transaction is committed. If necessary, the query waits for the results of forwarded write operations to be replicated to the current Region. It doesn't wait for updated results from write operations performed in other Regions or in other sessions within the current Region.

With the read consistency set to GLOBAL, a session in a secondary Amazon Region sees changes made by that session. It also sees all committed changes from both the primary Amazon Region and other secondary Amazon Regions. Each query might wait for a period that varies depending on the amount of session lag. The query proceeds when the secondary cluster is up-to-date with all committed data from the primary cluster, as of the time that the query began.

For more information about all the parameters involved with write forwarding, see Configuration parameters for write forwarding in Aurora MySQL.

Examples of using write forwarding

These examples use aurora_replica_read_consistency as a session variable. For Aurora MySQL version 3.04 and higher, you can use aurora_replica_read_consistency as either a session variable or as a DB cluster parameter.

In the following example, the primary cluster is in the US East (N. Virginia) Region. The secondary cluster is in the US East (Ohio) Region. The example shows the effects of running INSERT statements followed by SELECT statements. Depending on the value of the aurora_replica_read_consistency setting, the results might differ depending on the timing of the statements. To achieve higher consistency, you might wait briefly before issuing the SELECT statement. Or Aurora can automatically wait until the results finish replicating before proceeding with SELECT.

In this example, there is a read consistency setting of eventual. Running an INSERT statement immediately followed by a SELECT statement still returns the value of COUNT(*). This value reflects the number of rows before the new row is inserted. Running the SELECT again a short time later returns the updated row count. The SELECT statements don't wait.

mysql> set aurora_replica_read_consistency = 'eventual';
mysql> select count(*) from t1;
+----------+
| count(*) |
+----------+
|        5 |
+----------+
1 row in set (0.00 sec)
mysql> insert into t1 values (6); select count(*) from t1;
+----------+
| count(*) |
+----------+
|        5 |
+----------+
1 row in set (0.00 sec)
mysql> select count(*) from t1;
+----------+
| count(*) |
+----------+
|        6 |
+----------+
1 row in set (0.00 sec)

With a read consistency setting of session, a SELECT statement immediately after an INSERT waits until the changes from the INSERT statement are visible. Subsequent SELECT statements don't wait.

mysql> set aurora_replica_read_consistency = 'session';
mysql> select count(*) from t1;
+----------+
| count(*) |
+----------+
|        6 |
+----------+
1 row in set (0.01 sec)
mysql> insert into t1 values (6); select count(*) from t1; select count(*) from t1;
Query OK, 1 row affected (0.08 sec)
+----------+
| count(*) |
+----------+
|        7 |
+----------+
1 row in set (0.37 sec)
+----------+
| count(*) |
+----------+
|        7 |
+----------+
1 row in set (0.00 sec)

With the read consistency setting still set to session, introducing a brief wait after performing an INSERT statement makes the updated row count available by the time the next SELECT statement runs.

mysql> insert into t1 values (6); select sleep(2); select count(*) from t1;
Query OK, 1 row affected (0.07 sec)
+----------+
| sleep(2) |
+----------+
|        0 |
+----------+
1 row in set (2.01 sec)
+----------+
| count(*) |
+----------+
|        8 |
+----------+
1 row in set (0.00 sec)

With a read consistency setting of global, each SELECT statement waits to ensure that all data changes as of the start time of the statement are visible before performing the query. The amount of waiting for each SELECT statement varies, depending on the amount of replication lag between the primary and secondary clusters.

mysql> set aurora_replica_read_consistency = 'global';
mysql> select count(*) from t1;
+----------+
| count(*) |
+----------+
|        8 |
+----------+
1 row in set (0.75 sec)
mysql> select count(*) from t1;
+----------+
| count(*) |
+----------+
|        8 |
+----------+
1 row in set (0.37 sec)
mysql> select count(*) from t1;
+----------+
| count(*) |
+----------+
|        8 |
+----------+
1 row in set (0.66 sec)

Running multipart statements with write forwarding in Aurora MySQL

A DML statement might consist of multiple parts, such as a INSERT ... SELECT statement or a DELETE ... WHERE statement. In this case, the entire statement is forwarded to the primary cluster and run there.

Transactions with write forwarding in Aurora MySQL

Whether the transaction is forwarded to the primary cluster depends on the access mode of the transaction. You can specify the access mode for the transaction by using the SET TRANSACTION statement or the START TRANSACTION statement. You can also specify the transaction access mode by changing the value of the Aurora MySQL session variable tx_read_only. You can only change this session value while you're connected to a secondary cluster that has write forwarding enabled.

If a long-running transaction doesn't issue any statement for a substantial period of time, it might exceed the idle timeout period. This period has a default of one minute. You can increase it up to one day. A transaction that exceeds the idle timeout is canceled by the primary cluster. The next subsequent statement you submit receives a timeout error. Then Aurora rolls back the transaction.

This type of error can occur in other cases when write forwarding becomes unavailable. For example, Aurora cancels any transactions that use write forwarding if you restart the primary cluster or if you turn off the write forwarding configuration setting.

Configuration parameters for write forwarding in Aurora MySQL

The Aurora cluster parameter groups include settings for the write forwarding feature. Because these are cluster parameters, all DB instances in each cluster have the same values for these variables. Details about these parameters are summarized in the following table, with usage notes after the table.

To control incoming write requests from secondary clusters, use these settings on the primary cluster:

The aurora_replica_read_consistency parameter is a session-level parameter that enables write forwarding. You use it in each session. You can specify EVENTUAL, SESSION, or GLOBAL for read consistency level. To learn more about consistency levels, see Isolation and consistency for write forwarding in Aurora MySQL. The following rules apply to this parameter:

Amazon CloudWatch metrics for write forwarding in Aurora MySQL

The following Amazon CloudWatch metrics and Aurora MySQL status variables apply to the primary cluster when you use write forwarding on one or more secondary clusters. These metrics are all measured on the writer DB instance in the primary cluster.

The following CloudWatch metrics and Aurora MySQL status variables apply to each secondary cluster. These metrics are measured on each reader DB instance in a secondary cluster with write forwarding enabled.