Failover scenarios

Last Updated : Oct 06, 2016 |
Prolog information
Metadata
Keywords:
The following section displays the failover scenarios in geo redundancy:
  • When a Context Store cluster is down, requests that are routed to the cluster will be redirected to the other cluster. This might result in GET requests going to a cluster which does not contain that context. For example, the request will fail in the following sequence of events:
    1. Context A is saved to cluster 1.
    2. Cluster 1 goes down before context A has been replicated to cluster 1.
    3. The load balancer attempts to route the GET request for context A to cluster 2.
  • When a cluster is shut down and restarted, existing data on the active cluster is not replicated to the other cluster. For example, the following data is not replicated:
    1. Shut down cluster 1.
    2. Create new contexts on cluster 2 and update existing contexts on cluster 2.
    3. Restart cluster 1
    4. Only the newly created contexts on cluster 2 are replicated to cluster 1.