Using ClearDB High Availability SQL Routers (CDBRs)
1.0 CDBR Overview
ClearDB High Availability SQL Routers (or CDBRs) are custom built, intelligent traffic managers that closely watch the health and state of our clusters. In the event that a cluster node becomes unhealthy or fails, our CDBRs can automatically redirect your database connections to the secondary master, ensuring 100% uptime without you having to worry about failing databases or even reconfiguring your applications to point to an alternative database.
While services are being restored, the CDBRs in cooperation with our core systems will monitor the state of the node being restored. Once MySQL on the failed node has been restored, our CDBRs will only redirect your connections back to it once the cluster has synchronized to within a pre-configured delay tolerance value (typically zero). This ensures that your database reads are always consistent.
1.1 CDBR-Enhanced Cluster Configurations
A Basic Enhanced Cluster Configuration Example
CDBR-Enhanced cluster configurations include the addition of one or more CDBR systems to your cluster, thus increasing its durability by ensuring that the CDBR will not go down if the database node does. We also support configurations in which CDBRs are load-balanced in a pool to ensure that no single CDBR can bring down an entire cluster’s access. Figure 1.0 illustrates a basic CDBR-Enhanced ClearDB Cluster.
Figure 1.0 – An example of a basic CDBR-enhanced ClearDB cluster topology.
A Large Enhanced Cluster Configuration Example
Some businesses have large read requirements for decision support workloads, as well as moderate to high write requirements for transaction processing. Our Hybrid cluster configurations serve as basis for providing custom cluster solutions that provide large read pool services as well as multiple masters for the ultimate in scalability and performance. Figure 1.1 illustrates an example of such a larger configuration. One really compelling part about how the read pool CDBRs work is how they can automatically turn on and off load balancing, depending on the configured replication delay tolerance. So if your cluster is either completely in sync or is within an “acceptable delay tolerance range”, the read pool CDBR will enable you to take advantage of your entire read pool in both regions, enabling massive scalability*.
Figure 1.1 – An example of a large CDBR-enhanced ClearDB cluster topology.
1.2 CDBR Pricing & Availability
If you’re a shared database customer, chances are your database is already running through one of our CDBRs right now. If you are one of our dedicated database cluster customers and would like to add one or more CDBRs to your cluster, please contact us so that we can provide you with a custom quote for your specific needs.
* Additional bandwidth charges may apply. Contact us for details.