This is a high level procedure on performing a restore for an API Gateway Cluster
This is for an API Gateway cluster of 2 or more nodes.
1. Explicitly take the cluster out of the Load Balancer rotation so that the API Gateways are not taking message traffic during the restore procedure.
2. Stop the API Gateway service on all nodes of the cluster:
service ssg stop
3. Stop replication on both of the API Gateway database nodes, by running the following command from the root prompt of both database nodes:
4. Run ssgrestore.sh (with the appropriate flags and paths necessary) on the Primary database node.
Note: The appropriate ssgrestore.sh command should be determined based on need from the following page:
5. Run ssgrestore.sh (with the appropriate flags and paths necessary) on the Secondary database node.
6. If there are processing nodes in the cluster, run ssgrestore.sh on any processing nodes (with the appropriate flags and paths necessary).
7. Re-initialize replication between Primary and Secondary database nodes, clone the 'ssg' database from the Primary database node to the Secondary database node (dropping 'ssg' database on the Secondary database node). Use the following procedure:
8. Start the API Gateway service on all nodes of the cluster:
service ssg start
9. Add the nodes back into the Load Balancer rotation.