4.2 Deploy a Highly Available FEPCluster

In a highly available FEP cluster, load balancing is possible by distributing read queries to replica instances.

In addition, if the master instance fails, the user can switch to the replica instance immediately to localize the business interruption period.

In a highly available configuration, you can select the synchronization mode for the replica instance. Synchronous replication is recommended for systems that cannot tolerate data loss in the event of a master instance failure.

Because multiple instances are created in a highly available configuration, licenses are required for each.

To deploy a highly available FEPCluster in given namespace, follow these steps:

[Prerequisites]

If the FEP cluster is running in HA mode, the backup and archive WAL volumes must be configured with shared storage (NFS, etc.) that supports ReadWriteMany. See the Openshift documentation for instructions on setting up shared storage. Also, the reference procedure is described in "Appendix C Utilize Shared Storage", so please check if necessary.

If you do not have shared storage, you can remove the backup section and the backup and archive volume sections to disable the backup feature and deploy the FEP cluster.

Note

If you are deploying on a Kubernetes cluster, Refer to "Custom Resource Parameters" in the Reference to create and apply a yaml file.

It is the same as the procedure from step 1 to step 3 in "4.1 Deploying FEPCluster using Operator".
Instead of step 4 in "4.1 Deploying FEPCluster using Operator", change to the YAML view and specify '3' for the "instances" parameter of "fep" in "spec". Specify the storage class for the prepared shared storage for the backup and archive WAL volumes.
It is the same as the procedure from step 5 to step 10 in "4.1 Deploying FEPCluster using Operator".
Three pods deployed and ready for a highly available FEPCluster.

Information

You can determine whether the master or replica pod is the master or replica pod by issuing the following command:

$ oc get pod -L feprole
NAME                                                READY   STATUS      RESTARTS     AGE     FEPROLE
fep-ansible-operator-88f7fb4b-5jh85                  1/1     Running     0           24m
new-fep-sts-0                                        2/2     Running     0           17m     master
new-fep-sts-1                                        2/2     Running     0           15m     replica
new-fep-sts-2                                        2/2     Running     0           13m     replica