Identity ManagementPDF version

Restarting instances

If you have instances that have been stopped on the cloud provider side or ones that are recommended to be restarted due to high CPU and memory usage, you can restart them without the need to start the instance on the cloud provider side or having to perform a repair operation on the stopped instance.

If the restart operation is performed on an instance in running state, then the restart operation stops the instance first, and then starts it. If the restart operation is performed on an instance that is already in stopped state, only the start operation is performed.

You can perform the restart operation even if the instances are in an unhealthy state or even if the cluster is in an unhealthy state.

Required roles

  • Owner
  • EnvironmentAdmin
  • These commands restart the instances directly on the cloud provider side but the instance can still be in decommissioned state in Cloudera Manager. Therefore, you might need to manually recommission the instances in Cloudera Manager or wait for autoscaling to happen.
  • On the Cloudera Manager side, some services on the instance might become unhealthy. In the recovery flow of autoscaling they will be picked up for recovery but it is possible that they cannot be recovered. This is only true for the hostgroups where autoscaling is supported and an autoscaling policy is defined.

Use the following commands to restart instances:

cdp environments repair-freeipa \
--environment-name [ENVIRONMENT_NAME] \
--instances [INSTANCES ...] \
--repair-type REBOOT \
[--force] [--no-force]

Example:

cdp environments repair-freeipa \
--environment-name vatsal-az-mowdev \
--force
--instances vatsal-az-mowdev-freeipa107482m1-800661e5 vatsal-az-mowdev-freeipa107482m0-3857e6a1 \
--repair-type REBOOT