RDS for MySQL is a reliable and scalable cloud database service. You can deploy databases within minutes and stay focused on application development. Say goodbye to stressful database O&M once and for all.

Cloud is now standard as more enterprises build and deploy on the cloud. When it comes to running enterprise databases on the cloud, however, there is an increased risk of transaction data loss, and this risk, which affects data reliability and integrity, may deter some enterprises from moving their valuable data to the cloud.

Huawei Cloud RDS for MySQL helps you guard against transaction loss by employing a number of proven high-availability mechanisms, protecting the continuity of your business.

Huawei Cloud RDS for MySQL Delivers High Availability

Huawei Cloud RDS for MySQL delivers a high level of availability with a primary/standby DB instance pair and semi-synchronous replication between them, coupled by a number of read replicas for additional data durability. This deployment ensures fast service recovery in the case of a primary database breakdown while also guaranteeing high efficiency for the primary database, with zero transaction loss.

How Huawei Cloud RDS for MySQL Delivers That High Availability

l  Accurately determine the replication status of the primary/secondary pair at the time of a primary database breakdown

The control node (HA) and the standby database both store the latest replication status and timestamp of the primary database. By comparing such information separately stored by the control node and the standby database, RDS for MySQL is able to accurately determine the replication status at the time of the primary database breakdown.

l  Automatically restore services based on the replication status of the failed primary database

Semi-synchronous replication is, in fact, synchronous replication most of the time. In rare cases (for example, when large transactions are executed), synchronous replication switches to asynchronous replication and then automatically switches back when the conditions that triggered asynchronous replication no longer exist. Depending on the replication status at the time of a primary database breakdown, service recovery proceeds in one of the following ways:

1. In the case of a primary database breakdown while synchronous replication is on, if the primary database can be brought up successfully, no transaction data will be lost, and services can recover within seconds.

2. In the case of a primary database breakdown while synchronous replication is on, if the primary database cannot be brought up, services will be smoothly failed over to the standby database. In this case, no transaction data will be lost either, and services can also recover within seconds.

3. In the case of a primary database breakdown while asynchronous replication is performed, a switchover to the standby database is disallowed, but if the primary database can be brought up successfully, no transaction data will be lost, and services can recover in seconds.

4. In the case of a primary database breakdown while asynchronous replication is performed and a switchover to the standby database is disallowed, but where the primary database cannot be brought up, the system can restore the primary database using the most recent backup available. In this case, no transaction data will be lost, and services can recover in a matter of minutes.

In all four cases, Huawei Cloud RDS for MySQL is able to ensure zero transaction loss and guarantee service recovery within seconds (or minutes in extreme conditions), protecting the continuity of your business.