Connecting to multiple clusters

When you configure a VMware region, you can connect to multiple clusters defined in the same vCenter. Each cluster is set as a new availability zone in the same region.

Before you begin

Be familiar with VMware support in OpenStack by reading the related OpenStack documentation.

About this task

Because an OpenStack Nova Compute service can connect only to one cluster in OpenStack, you must create a new OpenStack Nova Compute service to connect to your new cluster. The new cluster is set as a new host aggregate in a new availability zone. In this procedure, the following names are used as examples:

new-cluster-availability-zone: Is the name of the new availability zone.
new-cluster-host-aggregate: Is the name of the new host aggregate.
nova-vmware.conf: Is the name of the Nova Compute service configuration file that is related to VMware, This name was specified when the VMware OpenStack Controller was installed.
openstack-nova-compute-vmware: Is the name of the Nova Compute service that is related to VMware. This name was specified when the VMware OpenStack Controller was installed.

Important: You must manually start, stop, or restart the newly created services, and must not use knife commands from the Deployment Server of IBM Cloud Manager with OpenStack.

Note:

You might add more cluster_name parameters in the Nova VMware configuration file instead of creating a new host aggregate and a new availability zone, but having two or more clusters in the same availability zone is not a good practice from a virtual machine placement point of view.
If you want to create several new OpenStack Nova Compute services to connect to different clusters, you can install new services on separate OpenStack Controllers to better manage your resource workload.

Procedure

Create the host aggregate and associate it with a new availability zone by running the following command:
```
nova aggregate-create new-cluster-host-aggregate new-cluster-availability-zone
```
This command also creates a new availability zone named new-cluster-availability-zone. Now you must create a new OpenStack Nova Compute service.
Create a copy of the Nova Compute service configuration file that is related to VMware, for example:
```
cp /etc/nova/nova-vmware.conf /etc/nova/nova-service-new-cluster.conf
```
Change the file ownership by running the following command:
```
chown nova:nova /etc/nova/nova-service-new-cluster.conf
```

Modify the /etc/nova/nova-service-new-cluster.conf file to set:

[DEFAULT]    
     default_availability_zone = new-cluster-availability-zone  
     default_schedule_zone = new-cluster-availability-zone
     storage_availability_zone = new-cluster-availability-zone
     host = new-cluster
     # Use a host name different from the VMware OpenStack Controller to 
     # avoid conflict with the first cluster configuration.
     
[vmware] 
     host_ip = <your vCenter IP address or host name>   
     cluster_name = <the name of the new cluster in your vCenter>

Add the host that you specified in step 3 for the new Compute service to the host aggregate by running the following command:

nova aggregate-add-host new-cluster-host-aggregate new-cluster

+----+----------------------------+-------------------------------+-------------
| Id | Name                       | Availability Zone             | Hosts       
+----+----------------------------+-------------------------------+-------------
| 2  | new-cluster-host-aggregate | new-cluster-availability-zone | new-cluster 
+----+----------------------------+-------------------------------+-------------
                                                                                
+---------------------------------------------------+                           
| Metadata                                          |                           
+---------------------------------------------------+                           
| 'availability_zone=new-cluster-availability-zone' |                           
+---------------------------------------------------+

Create copy of service file in /usr/lib/systemd/system.
For example, run the following command in a single line:
```
cp openstack-nova-compute-vcenter-Cluster1.service 
openstack-nova-compute-vcenter-Cluster2.service
```

Update the new service file with relevant values for the following parameters:

--config-file - Points to the new nova cluster configuration file.
--log-file - Points to the new nova log file.

Example:

[Unit]
Description=OpenStack Nova Compute Server for vcenter-Cluster2
After=syslog.target network.target

[Service]
Environment=LIBGUESTFS_ATTACH_METHOD=appliance
Restart=on-failure
User=nova
ExecStart=/usr/bin/nova-compute --config-file /etc/nova/nova-vcenter-Cluster2.conf 
--log-file /var/log/nova/nova-compute-vcenter-Cluster2.log

[Install]
WantedBy=multi-user.target

Run the following command in a single line to create a link to the new service file.

In -s  /usr/lib/systemd/system/openstack-nova-compute-vcenter-Cluster2.service   
/etc/systemd/system/multi-user.target.wants/openstack-nova-compute-vcenter-Cluster2.service

Change the permission on this file to 755.

chmod 755 openstack-nova-compute-vcenter-Cluster2.service

Run the following commands to start the services:

Under /usr/lib/systemd/system:

chkconfig openstack-nova-compute-vcenter-Cluster2 on
service  openstack-nova-compute-vcenter-Cluster2 start
service  openstack-nova-compute-vcenter-Cluster2 status

Verify that the new service is up and running. Run the following command to check whether the new availability zone that is named new-cluster-availability-zone is shown:

nova availability-zone-list

+---------------------------------------------+----------------------------------------+
| Name                                        | Status                                 | 
+---------------------------------------------+----------------------------------------+ 
| internal                                    | available                              | 
| |- <your-local-host-name>                   |                                        | 
| | |- nova-conductor                         | enabled :-) 2015-05-07T05:15:44.766879 | 
| | |- nova-vmware                            | enabled :-) 2015-05-07T05:15:51.017709 | 
| | |- nova-consoleauth                       | enabled :-) 2015-05-07T05:15:49.413705 | 
| | |- nova-cert                              | enabled :-) 2015-05-07T05:15:47.481551 | 
| | |- nova-scheduler                         | enabled :-) 2015-05-07T05:15:47.736521 | 
| nova                                        | available                              |  
| |- <your-local-host-name>                   |                                        | 
| | |- nova-compute                           | enabled :-) 2015-05-07T05:15:43.274219 |
| new-cluster-availability-zone               | available                              | 
| |- new-cluster                              |                                        | 
| | |- nova-compute                           | enabled :-) 2015-05-07T05:15:44.309888 | 
+---------------------------------------------+----------------------------------------+

For troubleshooting, see the /var/log/nova/compute-new-cluster.log file.

What to do next

After you configured your VMware region to connect to multiple clusters, you must run the vmware-discovery process as described in Configuring vmware-discovery. In the vmware-discovery configuration file, specify the name of your clusters, which are separated by a comma, in the clusters parameter. For example:

clusters = cluster1,cluster2,new-cluster