The system does not support power compute node deployment through the Web interface. You must convert VMware compute nodes into power compute nodes as follows:
If you have access to VMware, create a VMware compute node on VMware as follows:
Create virtualization settings.
Create a compute node.
For more information, see "Manage compute nodes."
If you do not have access to VMware, create a VMware compute node as follows:
Use an HTTP request tool to request token information. The IP address in the URL is the management IP address of the system. Please replace the IP address based on your environment.
Post http://172.25.51.97:8000/sys/identity/v1/tokens
Header
Accept:application/json
Content-Type:application/json
Body
{
"identity":
{"username":"admin",
"password":"xxx" */Password of the current administrator/*
}
}
Create a VMware compute node. The IP address in the URL is the management IP address of the system. Please replace the IP address based on your environment.
Post http://172.25.51.97:8000/os/compute/v1/xxos/computenode
Header:
Accept:application/json
Content-Type:application/json
X-Auto-Token: {Token obtained in the previous step}
Body:
{
"allPass" : "xxos",
"resourceNode" : "host",
"useLocalStorage" : "True",
"linkClone" : "False",
"mode" : "expand",
"ip" : "172.25.51.97", */System management IP address/*
"hostName" : "host818181", */Compute node name/*
"vmType" : 1, */VMware:1/*
"hostIp" : "172.25.17.181", */Compute node IP/*
"userName" : "[email protected]", */Compute node username/*
"password" : "vmware", */Compute node password/*
"poolName" : "xxos",
"vxlanOverlayMode" : "3", */Networking mode: VLAN—3, VXLAN host overlay—2, VXLAN network overlay—1/*
"storageZone" : "cinder818181",
"clusterName" : "xxos",
"clusterId" : "domain-c7",
"initMode" : "vmware_api",
"vswitch" : [ {
"netName" : "dmxhbg==", */Value encoded by Base64/*
"device" : "vSwitch0"
} ],
"pci" : [ ],
"vmwareVersion" : "2"
,"pci":[]}
Check the service status in the proper Nova container and verify that the pods named with the compute node name are running.
Log in to the control node through SSH and load environment variables.
source /opt/bin/common/tool.sh
Locate the Nova container ID.
docker ps | grep host name
Enter the Nova container.
docker exec –it container_id bash
Load environment variables.
source /root/admin-openrc.sh
View the service status.
nova service-list
· As a best practice, back up the configuration of the original environment to prevent misoperations from causing data loss. In the configuration files, the first letters of words are in lowercase. · Please obtain the newest power driver from your power service provider or copy it from a deployed power host. |
Log in to the system control node through SSH and upload the newest power driver file to /tmp.
Load environment variables.
[root@d013rc6-c203 ~]# source /opt/bin/common/tool.sh
Locate the container ID based on the host name of the compute node.
[root@d013rc4-c202 ~]# pod | grep com61
default com234rc-fht5n 1/1 Running 3 5d 10.101.29.28 172.25.48.203
Enter the compute node container.
[root@d013rc6-c203 ~]# docker exec -it container_id bash
Use a VI editor to add the following information to the /etc/nova.conf file of the compute node container. Please replace the URL, username, and password used for accessing the PowerCenter console according to your environment.
compute_driver = pcenter.Pcenter.Driver
[pcenter]
pcenter_url = http://lingcloudpower.ticp.net:25549/rest # Contact LingCloud Power to obtain the port number.##
pcenter_username = nova
pcenter_password = nova1234
group = cec
version = 1
Use a VI editor to add the following information to the /etc/cinder/cinder.conf file of the compute node container. Please replace the URL, username, and password used for accessing the PowerCenter console according to your environment.
enabled_backends = pcenter
[pcenter]
volume_group = pcenter
volume_driver = cinder.volumedrivers.pcenter.PcenterDriver
volume_backend_name = pcenter
pcenter_url = http://114.242.9.246:25549/rest
pcenter_name = cinder
pcenter_password = cinder1234
scheme = vglv
group = cec
version = 1
Use a VI editor to add the following information to the /etc/ceilometer.conf file of the compute node container. Please replace the URL, username, and password used for accessing the PowerCenter console according to your environment.
hypervisor_inspector = pcenter
[pcenter]
pcenter_url = http://114.242.9.246:25549/rest
pcenter_username = ceilometer
pcenter_password = ceilometer1234
version = 1
Use a VI editor to add the following information to the /usr/lib/python2.7/site-packages/ceilometer-9.0.1-py2.7-egg-info/entry_points.txt file of the compute node container.
pcenter = ceilometer:compute.virt.pcenter.inspector:PcenterInspector
Create folders to store driver files of PowerCenter in the compute node container.
Create a folder for Nova driver files.
[root@com61rc /]# mkdir /usr/lib/python2.7/site-packages/nova/virt/pcenter
Create a folder for Cinder driver files.
[root@com61rc /]# mkdir /usr/lib/python2.7/site-packages/cinder/volume/drivers/pcenter
Create a folder for Ceilometer driver files.
[root@com61rc /]# mkdir usr/lib/python2.7/site-packages/ceilometer/compute/virt/pcenter
Copy the Nova, Cinder, and Ceilometer driver files for PowerCenter to the proper folders in the compute node container.
[root@d013rc6-c203 ~]# docker cp /tmp/driver file name container_id:/target folder path
Restart the openstack-nova-compute, openstack-cinder-volume, and openstack-ceilometer-polling services for the compute node container.
[root@com61rc /]#service openstack-nova-compute restart
[root@com61rc /]#service openstack-cinder-volume restart
[root@com61rc /]#service openstack-ceilometer-polling restart
Verify that the openstack-nova-compute, openstack-cinder-volume, and openstack-ceilometer-polling services are running in the compute node container.
[root@com61rc /]#service openstack-nova-compute status
[root@com61rc /]#service openstack-cinder-volume status
[root@com61rc /]#service openstack-ceilometer-polling status