Restarting IBM Cloud Manager with OpenStack services

If you need to restart services for IBM Cloud Manager with OpenStack, use the IBM Cloud Manager with OpenStack services command and specify the restart action. You can use the command to restart IBM Cloud Manager with OpenStack services on a specific node or on all nodes in a topology.

About this task

Note: For an HA topology, restarting services on the controller nodes is more complicated because Pacemaker controls the services. For more information about managing services on HA controller nodes, see Managing OpenStack services on HA Controller Nodes.

To restart the IBM Cloud Manager with OpenStack services, run the following command for any node that is not a high availability (HA) controller node in your topology deployment:

$ knife os manage services restart –-node your-node-FQDN

All IBM Cloud Manager with OpenStack services on the node are restarted.

If you restart an OpenStack controller node, you should also restart the IBM Cloud Manager with OpenStack services on all of the OpenStack compute nodes in the topology. Issue the knife os manage services restart command for each OpenStack compute node to restart the services on each compute node in the topology.

Alternatively, run the following command to restart the IBM Cloud Manager with OpenStack services in the proper order for each node in a known topology by using a topology JSON file:

$ knife os manage services restart –-topology-file your-topology-name.json

All IBM Cloud Manager with OpenStack services on all nodes that are not HA controller nodes in the topology are restarted in the proper order.

By default, the restart action of the IBM Cloud Manager with OpenStack services command does not affect the HA controller nodes that might exist in a topology. If you have a HA controller + n compute topology, the IBM Cloud Manager with OpenStack services on the HA controller nodes are managed by HA components. The management of the OpenStack services by the HA components includes monitoring and restarting services as needed to maintain availability of the OpenStack services. This essentially eliminates the need to manually restart the OpenStack services on HA controller nodes.

In certain cases it might be desirable to restart an OpenStack service on one or all HA controller nodes. You have several options to restart the OpenStack service:

The first way is to disable and enable a resource.
1. Use the following command to shut down the service on any node that it is running on.
```
pcs resource disable resource-name --wait=wait_time_in_seconds
```
2. Start the service again by running the following command. Pacemaker starts the service on the appropriate nodes. A reasonable wait time is 450 seconds.
```
pcs resource enable resource-name --wait=wait_time_in_seconds
```
The other way to restart an OpenStack service on a specific system is to kill a process on the node that you choose and let Pacemaker restart it. Pacemaker either restarts the service on the same node, or moves it to a different node, depending on its determination of where best to run the service. Use the ps linux command to find the process, and then use the kill linux command to kill the process or processes. Some services have multiple processes. Many processes have a parent process that controls a set of child processes. Usually it is sufficient to kill the parent process; the child processes end simultaneously.
Use the following command to clean up the service on all nodes. This will reset the service's Pacemaker resource status and failure count and allow Pacemaker to re-detect the current state of the service. This is an advanced option and should only be used if the previous options do not attempt to restart the service.
```
pcs resource cleanup resource-name
```