From Fedora Project Wiki

m (Syntax in 13.2 was wrong)
No edit summary
Line 84: Line 84:


12. Create a keystone tenant and user.
12. Create a keystone tenant and user.
12.1 Create the keystone tenant
    keystone tenant-list    # Record 'services' tenant ID
    keystone user-create --name=quantum --pass="mypass" --email=demo@example.com --tenant-id <services_ID>
12.2 Add the quantum user to the admin role
    keystone role-list      # Record 'admin' role
    keystone user-list      # Record 'quantum' user ID
    keystone user-role-add --user_id <quantum_ID> --role_id <admin_ID> --tenant_id <services_ID>


13. Now one is able to start to use the Quantum CLI
13. Now one is able to start to use the Quantum CLI

Revision as of 19:33, 4 March 2013

The following is a description of how to go from packstack with traditional nova networking to Quantum. In the example below the OpenvSwitch plugin is used:

1. Terminate nova networking

   service openstack-nova-network stop
   chkconfig openstack-nova-network off

2. Install quantum service

   yum install openstack-quantum

3. Install OpenvSwicth plugin

   yum install openstack-quantum-openvswitch

4. The Quantum installation scripts make use of keystone environment variables. Make sure that the administrator environment variables are set.

   source keystonerc_admin

5. Configure quantum service

   quantum-server-setup

6.1. The above script will ask if the nova parameters need to be updated. Restart the nova compute service. The script will configure Quantum as the networking module for Nova.

   service openstack-nova-compute restart

6.2 Start the openvswitch service

   service openvswitch start
   chkconfig openvswitch on

6.3 Create the integration bridge (for the private network)

   ovs-vsctl add-br br-int

6.4 Create the externak bridge (for the layer 3 agent)

   ovs-vsctl add-br br-ex

6.5 Ensure that the interfaces are persistant after reboot. This is done by creating interface files, for example for br-int:

   DEVICE=br-int
   DEVICETYPE=ovs
   TYPE=OVSBridge
   ONBOOT=yes

7. Start quantum service

   service quantum-server start
   chkconfig quantum-server on

7.1 Start quantum agent

   service quantum-openvswitch-agent start
   chkconfig quantum-openvswitch-agent on

8. In order for the IPAM to take place one needs to invoke the DHCP agent. Use the DHCP setup tool

   quantum-dhcp-setup

8.1. Start DHCP agent

   service quantum-dhcp-agent start
   chkconfg quantum-dhcp-agent on

9. In order to support floating IP's one need to invoke the Layer 3 agent. Use the L3 agent setup tool

   quantum-l3-setup

9.1. Start the L3 agent

   service quantum-l3-agent start
   chkconfig quantum-l3-agent on 

10. Ensure that the openvswitch cleanup utility is enabled (this is used after reboot to ensure that device management of the quantum agents is not affected by the fact that the openvswitch creates the interfaces at boot)

   chkconfig quantum-ovs-cleanup on

11. Create a Quantum endpoint with keystone

   keystone service-create --name=quantum --type=network --description="Quantum Service"
   keystone endpoint-create --region RegionOne --service-id<ID>  --publicurl "http://127.0.0.1:9696" --adminurl "http://127.0.0.1:9696" --internalurl "http://127.0.0.1:9696"

12. Create a keystone tenant and user.

12.1 Create the keystone tenant

   keystone tenant-list     # Record 'services' tenant ID
   keystone user-create --name=quantum --pass="mypass" --email=demo@example.com --tenant-id <services_ID>

12.2 Add the quantum user to the admin role

   keystone role-list       # Record 'admin' role
   keystone user-list       # Record 'quantum' user ID
   keystone user-role-add --user_id <quantum_ID> --role_id <admin_ID> --tenant_id <services_ID>

13. Now one is able to start to use the Quantum CLI

13.1 Create a private network

   quantum net-create private

13.2 Create a subnet

   quantum subnet-create private 10.0.0.0/24

14. Validate that a port has been created by the DHCP agent (quantum port-list). This port will have IP address 10.0.0.2. The gateway will 10.0.0.1.

15. At this stage VM's can be deployed and they will receive IP addresses from the DHCP service.