The following command can be used to roll a cluster. Rolling a cluster will gradually replace the instances of the cluster with new instances. This is a useful action to apply cluster configuration changes with immediate effect.


ROLL CLUSTER <cluster_name> [ NO_GRACE_PERIOD = { TRUE | FALSE } ]


  • NO_GRACE_PERIOD - If TRUE, the cluster's instances will shut down immediately with no grace period. This is in contrast to the gradual roll of instances applied by default. Using this option can cause some downtime of the web UI if performed on the API cluster.

Last updated