You can use this procedure to make the current standby VSD cluster the active VSD cluster. You can promote the standby cluster to active when both the active and standby clusters are in good health. Using the vsd_force_cluser_failover flag (see below), you can promote the standby cluster even when the primary cluster is unhealthy or unreachable.
When you use MetroAE to execute the failover procedure, the vsds.yml file in your deployment must be configured with data for 6 VSDs. The first 3 VSDs (VSDs 1-3 in the vsds.yml file) must correspond to the 3 VSDs that are currently active VSD cluster and the second 3 VSDs (VSDs 4-6 in the vsds.yml file) must correspond to the 3 VSDs that are currently standby. You can create a new deployment for the failover operation or you can re-use the same deployment that was used by MetroAE to install the active/standby cluster in the first place. It is not required that MetroAE be used to install the active/standby clusters in the first place.
Note that the deployment must be configured for the current active/standby situation. You can use this procedure to switch back and forth (failover and failback) your clusters, but the order of VSDs in vsds.yml must be changed between runs or you can create separate deployments: One for deactivating the active cluster and promoting the standby cluster to active and one for reversing the operation.
When your deployment is correct, use the following command to deactivate the current active cluster and promote the standby cluster to active:
metroae vsd failover [deployment_name]
If both clusters are in good shape, your failover will complete successfully. If the current primary cluster is in poor health or unreachable, MetroAE will print an error and quit. If you want to ignore these errors and force the current standby cluster to become primary, use following command:
metroae vsd failover [deployment_name] -e vsd_force_cluster_failover=yes
MetroAE will also execute a health check on the current standby cluster. If the health check fails on the standby cluster, MetroAE will print an error and quit. If you want to ignore these errors and promote the standby cluster anyway, use the following command:
metroae vsd failover [deployment_name] -e skip_health_check=true
Note that you can use both vsd_force_cluster_failover and skip_health_check on one command:
metroae vsd failover [deployment_name] -e vsd_force_cluster_failover=yes -e skip_health_check=true
Get support via the forums on the MetroAE site.
Ask questions and contact us directly at devops@nuagenetworks.net.
Report bugs you find and suggest new features and enhancements via the GitHub Issues feature.
You may also contribute to MetroAE by submitting your own code to the project.