Red Hat Training

A Red Hat training course is available for Red Hat OpenStack Platform

Chapter 5. Test the deployment

5.1. Perform a basic test

The basic test will verify that instances are able to ping each other. It will also check the Floating IP SSH access. This example describes how you can perform the test from the undercloud.

This procedure requires you to follow a large number of individual steps; for convenience, the procedure was divided into smaller parts. However, the steps must be followed in the given order.

Note

In this setup, a flat network is used to create the External network, and VXLAN is used for the Tenant networks. VLAN External networks and VLAN Tenant networks are also supported, depending on the desired deployment.

5.1.1. Create a new network for testing

1. Source the credentials to access the overcloud:

$ source /home/stack/overcloudrc

2. Create an external neutron network that will be used to access the instance from outside of the overcloud:

$ openstack network create --external --project $(openstack project show service | grep id | awk '{ print $4 }')  --provider-network-type flat --provider-physical-network external

3. Create the corresponding neutron subnet for the new external network (created in the previous step):

$ openstack subnet create  --project $(openstack project show service | grep id | awk '{ print $4 }') --no-dhcp --network external --gateway 192.168.37.1 --allocation-pool start=192.168.37.200,end=192.168.37.220 --subnet-range 192.168.37.0/24 external-subnet

4. Download the cirros image to be used for creating overcloud instances:

$ wget http://download.cirros-cloud.net/0.3.4/cirros-0.3.4-x86_64-disk.img

5. Upload the cirros image into glance on the overcloud:

$ openstack image create cirros --public --file ./cirros-0.3.4-x86_64-disk.img --disk-format qcow2 --container-format bare

6. Create a tiny flavor to use for overcloud instances:

$ openstack flavor create m1.tiny --ram 512 --disk 1 --public

7. Create a VXLAN-based tenant network to host the instances:

$ openstack network create net_test --provider-network-type=vxlan --provider-segment 100

8. Create a subnet for the tenant network (created in the previous step):

$ openstack subnet create --network net_test --subnet-range 123.123.123.0/24 test

9. Find and store the ID of the tenant network:

$ net_mgmt_id=$(openstack network list | grep net_test | awk '{print $2}')

10. Create an instance called cirros1 and attach it to the net_test network:

$ openstack server create --flavor m1.tiny --image cirros --nic net-id=$net_mgmt_id cirros1

11. Create a second instance called cirros2, also attached to the net_test network:

$ openstack server create --flavor m1.tiny --image cirros --nic net-id=$net_mgmt_id cirros2

5.1.2. Set up networking in the test environment

1. Find and store the ID of the admin project:

$ admin_project_id=$(openstack project list | grep admin | awk '{print $2}')

2. Find and store the admin project’s default security group:

$ admin_sec_group_id=$(openstack security group list | grep $admin_project_id | awk '{print $2}')

3. Add a rule to the admin default security group to allow ICMP traffic ingress:

$ openstack security group rule create $admin_sec_group_id --protocol icmp --ingress

4. Add a rule to the admin default security group to allow ICMP traffic egress:

$ openstack security group rule create $admin_sec_group_id --protocol icmp --egress

5. Add a rule to the admin default security group to allow SSH traffic ingress:

$ openstack security group rule create $admin_sec_group_id --protocol tcp --dst-port 22 --ingress

6. Add a rule to the admin default security group to allow SSH traffic egress:

$ openstack security group rule create $admin_sec_group_id --protocol tcp --dst-port 22 --egress

5.1.3. Test the connectivity

1. From horizon, you should be able to access the novnc console for an instance. Use the password from overcloudrc to login to horizon as admin. The default login for cirros images is the username cirros, and cubswin:) as the password.

2. From the novnc console, verify that the instance received a DHCP address:

$ ip addr show
Note

Another method of doing this is by using the nova console-log <instance id> from the undercloud, which will show if a DHCP lease was obtained.

3. Now repeat the steps 1 and 2 for all other instances.

4. From one instance, attempt to ping the other instances. This will validate the basic Tenant network connectivity in the overcloud.

5. Verify that you can reach other instances by using a Floating IP.

5.1.4. Create devices

1. Create a floating IP on the external network to be associated with cirros1 instance:

$ openstack floating ip create external

2. Create a router which will be used to handle NAT between the floating IP and cirros1 tenant IP:

$ openstack router create test

3. Set the gateway of the router to be the external network:

$ neutron router-gateway-set test external

4. Add and interface to the router attached to the tenant network:

$ neutron router-interface-add test test

5. Find and store the floating IP created in Step 23:

floating_ip=$(openstack floating ip list | head -n -1 | grep -Eo '[0-9]+\.[0-9]+\.[0-9]+\.[0-9]+')

6. Associate the floating IP with the cirros1 instance:

$ openstack server add floating ip cirros1 $floating_ip

7. From a node that has external network access, attempt to login to the instance:

$ ssh cirros@$floating_ip

5.2. Perform advanced tests

Several components of the OpenDaylight configuration and deployment may be checked post deployment. To test specific parts of the installation, you need to follow several procedures. Each procedure is described separately.

The procedures are to be performed on the overcloud nodes.

5.2.1. Connect to overcloud nodes

  1. Login onto the undercloud.
  2. Run the following command to start the process:

    $ source /home/stack/stackrc
  3. List all instances:

    $ nova list
  4. Choose the required instance and note its IP address in the list:
  5. Connect to the machine. You will use the IP address from the list above:

    $ ssh heat-admin@<IP from step 3>
  6. Switch to superuser:

    $ sudo -i

5.2.2. Test OpenDaylight

To test that OpenDaylight is working, you have to verify that the service is up and that the particular features are correctly loaded.

  1. As a superuser, login to the node.
  2. Verify that the OpenDaylight service is active:

    # systemctl status opendaylight
  3. Verify that HAProxy is properly configured to listen on port 8081:

    # grep -A7 opendaylight /etc/haproxy/haproxy.cfg
  4. Connect to the karaf account:

    $ ssh -p 8101 karaf@localhost
  5. List the installed features.

    $ feature:list | grep odl-netvirt-openstack
  6. Verify that the API is up and running.

    # web:list | grep neutron
  7. Verify that inter-nodes' VXLAN tunnels are up.

    # vxlan:show
  8. To test that the REST API is responding correctly, you can list the modules that are using it.

    # curl -u "admin:admin" http://localhost:8181/restconf/modules

    The output will be similar (the example has been shortened).

    {"modules":{"module":[{"name":"netty-event-executor","revision":"2013-11-12","namespace":"urn:opendaylight:params:xml:ns:yang:controller:netty:eventexecutor"},{"name" ...
  9. You can list the REST streams.

    # curl -u "admin:admin" http://localhost:8181/restconf/streams

    You will get something like this:

    {"streams":{}}
  10. Enter the following command to verify that NetVirt is ready and running.

    # curl -u "admin:admin" http://localhost:8181/restconf/operational/network-topology:network-topology/topology/netvirt:1

    The following output will confirm it.

    {"topology":[{"topology-id":"netvirt:1"}]}

More information

  • Step 3: As mentioned before, OpenDaylight is not running in HA mode yet, therefore the service is only active on one node.
  • Step 5: If there is an x in the third column of the list, as generated during the procedure, then the feature is correctly installed.
  • Step 6: This API endpoint is set in /etc/neutron/plugins/ml2/ml2_conf.ini and used by the neutron to communicate with OpenDaylight.

5.2.3. Test Open vSwitch

In order to validate Open vSwitch, connect to one of the Compute nodes and verify that it is properly configured and connected to OpenDaylight.

  1. Connect to one of the Compute nodes in the overcloud as a superuser.
  2. List the Open vSwitch settings.

    # ovs-vsctl show
  3. Notice multiple Managers in the output (lines 2 and 3 in the example).

    4b624d8f-a7af-4f0f-b56a-b8cfabf7635d
        Manager "ptcp:6639:127.0.0.1"
        Manager "tcp:192.0.2.4:6640"
            is_connected: true
        Bridge br-extu
            Port br-ex
                Interface br-ex
                    type: internal
            Port "eth2"
                Interface "eth2"
            Port br-ex-int-patch
                Interface br-ex-int-patch
                    type: patch
                    options: {peer=br-ex-patch}
        Bridge br-int
            Controller "tcp:192.0.2.4:6653"
                is_connected: true
            fail_mode: secure
            Port br-int
                Interface br-int
                    type: internal
            Port br-ex-patch
                Interface br-ex-patch
                    type: patch
                    options: {peer=br-ex-int-patch}
        ovs_version: "2.5.0"
  4. Verify that the tcp manager points to the IP of the node where OpenDaylight is running.
  5. Verify that the Managers show is_connected: true to ensure that connectivity to OpenDaylight from OVS is established and uses the OVSDB protocol.
  6. Verify that each bridge (other than br-int) exists and matches the NIC template used for deployment with the Compute role.
  7. Verify that the tcp connection corresponds to the IP where the OpenDaylight service is running.
  8. Verify that the bridge br-int shows is_connected: true and an OpenFlow protocol connection to OpenDaylight is established.

More information

  • The br-int bridge is created automatically by OpenDaylight.

5.2.4. Verify the Open vSwitch configuration on Compute nodes.

  1. Connect to a Compute node as a superuser.
  2. List the Open vSwitch configuration settings.

    # ovs-vsctl list open_vswitch
  3. Read the output. It will be similar to this example.

    _uuid               : 4b624d8f-a7af-4f0f-b56a-b8cfabf7635d
    bridges             : [11127421-3bcc-4f9a-9040-ff8b88486508, 350135a4-4627-4e1b-8bef-56a1e4249bef]
    cur_cfg             : 7
    datapath_types      : [netdev, system]
    db_version          : "7.12.1"
    external_ids        : {system-id="b8d16d0b-a40a-47c8-a767-e118fe22759e"}
    iface_types         : [geneve, gre, internal, ipsec_gre, lisp, patch, stt, system, tap, vxlan]
    manager_options     : [c66f2e87-4724-448a-b9df-837d56b9f4a9, defec179-720e-458e-8875-ea763a0d8909]
    next_cfg            : 7
    other_config        : {local_ip="11.0.0.30", provider_mappings="datacentre:br-ex"}
    ovs_version         : "2.5.0"
    ssl                 : []
    statistics          : {}
    system_type         : RedHatEnterpriseServer
    system_version      : "7.3-Maipo"
  1. Verify that the value of the other_config option has the correct local_ip set for the local interface that connects to the Tenant network through VXLAN tunnels.
  2. Verify that the provider_mappings value under the other_config option matches the value given in the OpenDaylightProviderMappings heat template parameter. This configuration maps the neutron logical networks to corresponding physical interfaces.

5.2.5. Verify neutron configuration

  1. Connect to the superuser account on one of the controller role nodes.
  2. Make sure that the file /etc/neutron/neutron.conf contains service_plugins=odl-router_v2.
  3. Check that the file /etc/neutron/plugin.ini contains the following ml2 configuration:

    [ml2]
    mechanism_drivers=opendaylight_v2
    
    [ml2_odl]
    password=admin
    username=admin
    url=http://192.0.2.9:8081/controller/nb/v2/neutron
  4. On one of the overcloud controllers, verify that neutron agents are running properly.

    # openstack network agent list
  5. Verify that both the Metadata and DHCP agents are in the up state (the admin_state_up option is True):

    +--------------------------------------+----------------+--------------------------+-------------------+-------+----------------+------------------------+
    | id                                   | agent_type     | host                     | availability_zone | alive | admin_state_up | binary                 |
    +--------------------------------------+----------------+--------------------------+-------------------+-------+----------------+------------------------+
    | 3be198c5-b3aa-4d0e-abb4-51b29db3af47 | Metadata agent | controller-0.localdomain |                   | :-)   | True           | neutron-metadata-agent |
    | 79579d47-dd7d-4ef3-9614-cd2f736043f3 | DHCP agent     | controller-0.localdomain | nova              | :-)   | True           | neutron-dhcp-agent     |
    +--------------------------------------+----------------+--------------------------+-------------------+-------+----------------+------------------------+

More information

  • The IP in the plugin.ini, mentioned in step 3, should be the InternalAPI Virtual IP Address (VIP).
  • Note, that there is no Open vSwitch agent, nor L3 agent, listed in output of step 5, which is a desired state, as both are now managed by OpenDaylight.