Disaster Recovery
The following clusters are supported by Spektra for Disaster Recovery as shown:
Source
Target
UA
UE-GCP
UE-AWS
UE-AWS
UE-AWS
UE-GCP
UE-GCP
UE-AWS
UA
UE-GCP
UA
If there’s an application running on one cluster (source) and you set up Disaster Recovery for it while specifying the target cluster on which you want to store the recovered application, the system moves data from the source cluster to the target cluster based on the replication interval you specify. Any data that changes in the duration on the source is moved to the target.
Replication is the process of copying data and meta-data from a primary site to a secondary site. Replication provides an extra measure of redundancy that can be invaluable if the primary site fails. Access (based on the duration you specify) to the replicated data minimizes downtime and its associated costs. This can be invaluable if the appliance disk cache or the entire appliance fails.
Replication streamlines disaster recovery processes by generating duplicate copies of all changes regularly. It can also simplify recovery from all types of disasters, including fire, flood, hurricane, virus, or worm.
Note
The Disaster Recovery feature is not supported for cloud clusters (GKE and EKS).
Pre-requisite
Create a project using at least two clusters.
Enabling DR between two Clusters
To enable DR between two clusters:
The application must be installed on the UA or UE cluster and should be in the running state.
On the Catalog page, select the work template to enable the DR.
You are directed to the installation tab of the workload.
Select the DR Settings tab and then select Update DR Settings.
Enable the DR toggle button.
Select a Source Network. This is the network on the cluster used for replication.
Select a Target Cluster from the drop-down. It can be either UA or UE. It should belong to the same project.
Select a Target Network for the Disaster Recovery process.
Select a Replication Period from the drop-down.
Select the toggle to retain the volume on deletion
Note
When DR UE (AWS) to UA or DR UE (GCP) to UA is performed, the replication agent pod on the source UE (AWS/GCP) cluster cannot communicate with the replication agent pod on the target UA cluster. Since UA clusters usually have private IP addresses, pods created on UA clusters get private IP addresses. DR will work, if the system administrator maps a public IP address or external IP address to a private endpoint of the agent pod.
Note
Ensure you run the diamanti-object-migration.sh –all script from node to view the available network on all applicable clusters. Other cases will still allow you to enable or disable Disaster Recovery, but the available networks will not appear.
Select Update.
DR has been enabled, and you can see the details on the DR settings tab.
Updating DR Settings
After you enable DR, update DR settings, where you can change the replication period and suspend replication:
Select the workload from the Catalog page.
Select the DR Settings tab.
Select Update DR Settings to display the DR Settings page.
Change the Replication Period or Suspend Replication if necessary.
Select Update.
Enabling Fire Drill
After you enable DR, you can start the fire-drill:
Select the workload from the catalog page.
Select the DR Settings tab.
Select the ellipsis next to the update DR settings.
Select Fire drill to display the Fire drill setting page.
Select Start Fire-drill to enable the application to Fire-drill from the source cluster to the target cluster.
Enabling Failover
After you enable DR, you can start the failover.
Important
Failover can only be enabled after replication has completed its initial sync.
Select the workload from the catalog page.
Select the DR Settings tab.
Select the ellipsis next to the update DR settings.
Select Fail Over to display the Fail Over setting page.
Select Fail Over to enable the application to failover from the source cluster to the target cluster.
Note
Wait for a while for the message Disaster Recovery PendingFailoverPhase to disappear after we initiate the failover.
Disabling DR
Once the DR is enabled, you can disable
Select the workload from the catalog page.
Select the DR Settings tab.
Select the update DR settings.
Disable the toggle.
Select Update to disable the DR for that installation.