Jump to contentJump to page navigation: previous page [access key p]/next page [access key n]
Applies to SUSE OpenStack Cloud 8

15 Installing ESX Computes and OVSvAPP Edit source

This section describes the installation step requirements for ESX Computes (nova-proxy) and OVSvAPP.

15.1 Before You Start Edit source

  1. Review the Chapter 2, Pre-Installation Checklist about recommended pre-installation tasks.

  2. Prepare the Cloud Lifecycle Manager node. The Cloud Lifecycle Manager must be accessible either directly or via ssh, and have SUSE Linux Enterprise Server 12 SP3 installed. All nodes must be accessible to the Cloud Lifecycle Manager. If the nodes do not have direct access to online Cloud subscription channels, the Cloud Lifecycle Manager node will need to host the Cloud repositories.

    1. If you followed the installation instructions for Cloud Lifecycle Manager server (see Chapter 3, Installing the Cloud Lifecycle Manager server), SUSE OpenStack Cloud software should already be installed. Double-check whether SUSE Linux Enterprise and SUSE OpenStack Cloud are properly registered at the SUSE Customer Center by starting YaST and running Software › Product Registration.

      If you have not yet installed SUSE OpenStack Cloud, do so by starting YaST and running Software › Product Registration › Select Extensions. Choose SUSE OpenStack Cloud and follow the on-screen instructions. Make sure to register SUSE OpenStack Cloud during the installation process and to install the software pattern patterns-cloud-ardana.

      tux > sudo zypper -n in patterns-cloud-ardana
    2. Ensure the SUSE OpenStack Cloud media repositories and updates repositories are made available to all nodes in your deployment. This can be accomplished either by configuring the Cloud Lifecycle Manager server as an SMT mirror as described in Chapter 4, Installing and Setting Up an SMT Server on the Cloud Lifecycle Manager server (Optional) or by syncing or mounting the Cloud and updates repositories to the Cloud Lifecycle Manager server as described in Chapter 5, Software Repository Setup.

    3. Configure passwordless sudo for the user created when setting up the node (as described in Section 3.4, “Creating a User”). Note that this is not the user ardana that will be used later in this procedure. In the following we assume you named the user cloud. Run the command visudo as user root and add the following line to the end of the file:

      CLOUD ALL = (root) NOPASSWD:ALL

      Make sure to replace CLOUD with your user name choice.

    4. Set the password for the user ardana:

      tux > sudo passwd ardana
    5. Become the user ardana:

      tux > su - ardana
    6. Place a copy of the SUSE Linux Enterprise Server 12 SP3 .iso in the ardana home directory, var/lib/ardana, and rename it to sles12sp3.iso.

    7. Install the templates, examples, and working model directories:

      ardana > /usr/bin/ardana-init

15.2 Setting Up the Cloud Lifecycle Manager Edit source

15.2.1 Installing the Cloud Lifecycle Manager Edit source

Running the ARDANA_INIT_AUTO=1 command is optional to avoid stopping for authentication at any step. You can also run ardana-initto launch the Cloud Lifecycle Manager. You will be prompted to enter an optional SSH passphrase, which is used to protect the key used by Ansible when connecting to its client nodes. If you do not want to use a passphrase, press Enter at the prompt.

If you have protected the SSH key with a passphrase, you can avoid having to enter the passphrase on every attempt by Ansible to connect to its client nodes with the following commands:

ardana > eval $(ssh-agent)
ardana > ssh-add ~/.ssh/id_rsa

The Cloud Lifecycle Manager will contain the installation scripts and configuration files to deploy your cloud. You can set up the Cloud Lifecycle Manager on a dedicated node or you do so on your first controller node. The default choice is to use the first controller node as the Cloud Lifecycle Manager.

  1. Download the product from:

    1. SUSE Customer Center

  2. Boot your Cloud Lifecycle Manager from the SLES ISO contained in the download.

  3. Enter install (all lower-case, exactly as spelled out here) to start installation.

  4. Select the language. Note that only the English language selection is currently supported.

  5. Select the location.

  6. Select the keyboard layout.

  7. Select the primary network interface, if prompted:

    1. Assign IP address, subnet mask, and default gateway

  8. Create new account:

    1. Enter a username.

    2. Enter a password.

    3. Enter time zone.

Once the initial installation is finished, complete the Cloud Lifecycle Manager setup with these steps:

  1. Ensure your Cloud Lifecycle Manager has a valid DNS nameserver specified in /etc/resolv.conf.

  2. Set the environment variable LC_ALL:

    export LC_ALL=C
    Note
    Note

    This can be added to ~/.bashrc or /etc/bash.bashrc.

The node should now have a working SLES setup.

15.3 Overview of ESXi and OVSvApp Edit source

ESXi is a hypervisor developed by VMware for deploying and serving virtual computers. OVSvApp is a service VM that allows for leveraging advanced networking capabilities that OpenStack Neutron provides. As a result, OpenStack features can be added quickly with minimum effort where ESXi is used. OVSvApp allows for hosting VMs on ESXi hypervisors together with the flexibility of creating port groups dynamically on Distributed Virtual Switches (DVS). Network traffic can then be steered through the OVSvApp VM which provides VLAN and VXLAN underlying infrastructure for VM communication and security features based on OpenStack. More information is available at the OpenStack wiki.

The diagram below illustrates the OVSvApp architecture.

15.4 VM Appliances Used in OVSvApp Implementation Edit source

The default configuration deployed with the Cloud Lifecycle Manager for VMware ESX hosts uses service appliances that run as VMs on the VMware hypervisor. There is one OVSvApp VM per VMware ESX host and one nova Compute Proxy per VMware cluster or VMware vCenter Server. Instructions for how to create a template for the Nova Compute Proxy or ovsvapp can be found at Section 15.9, “Create a SUSE-based Virtual Appliance Template in vCenter”.

15.4.1 OVSvApp VM Edit source

OVSvApp implementation is comprised of:

  • a service VM called OVSvApp VM hosted on each ESXi hypervisor within a cluster, and

  • two vSphere Distributed vSwitches (DVS).

OVSvApp VMs run SUSE Linux Enterprise and have Open vSwitch installed with an agent called OVSvApp agent. The OVSvApp VM routes network traffic to the various VMware tenants and cooperates with the OpenStack deployment to configure the appropriate port and network settings for VMware tenants.

15.4.2 Nova Compute Proxy VM Edit source

The Nova compute proxy is the nova-compute service for VMware ESX. Only one instance of this service is required for each ESX cluster that is deployed and is communicating with a single VMware vCenter server. (This is not like KVM where the nova-compute service must run on every KVM Host.) The single instance of nova-compute service can run in the OpenStack controller node or any other service node in your cloud. The main component of the nova-compute VM is the OVSvApp nova VCDriver that talks to the VMware vCenter server to perform VM operations such as VM creation and deletion.

15.5 Prerequisites for Installing ESXi and Managing with vCenter Edit source

ESX/vCenter integration is not fully automatic. vCenter administrators are responsible for taking steps to ensure secure operation.

  • The VMware administrator is responsible for administration of the vCenter servers and the ESX nodes using the VMware administration tools. These responsibilities include:

    • Installing and configuring vCenter server

    • Installing and configuring ESX server and ESX cluster

    • Installing and configuring shared datastores

    • Establishing network connectivity between the ESX network and the Cloud Lifecycle Manager OpenStack management network

  • The VMware administration staff is responsible for the review of vCenter logs. These logs are not automatically included in Cloud Lifecycle Manager OpenStack centralized logging.

  • The VMware administrator is responsible for administration of the vCenter servers and the ESX nodes using the VMware administration tools.

  • Logging levels for vCenter should be set appropriately to prevent logging of the password for the Cloud Lifecycle Manager OpenStack message queue.

  • The vCenter cluster and ESX Compute nodes must be appropriately backed up.

  • Backup procedures for vCenter should ensure that the file containing the Cloud Lifecycle Manager OpenStack configuration as part of Nova and Cinder volume services is backed up and the backups are protected appropriately.

  • Since the file containing the Cloud Lifecycle Manager OpenStack message queue password could appear in the swap area of a vCenter server, appropriate controls should be applied to the vCenter cluster to prevent discovery of the password via snooping of the swap area or memory dumps.

  • It is recommended to have a common shared storage for all the ESXi hosts in a particular cluster.

  • Ensure that you have enabled HA (High Availability) and DRS (Distributed Resource Scheduler) settings in a cluster configuration before running the installation. DRS and HA are disabled only for OVSvApp. This is done so that it does not move to a different host. If you do not enable DRS and HA prior to installation then you will not be able to disable it only for OVSvApp. As a result DRS or HA could migrate OVSvApp to a different host, which would create a network loop.

Note
Note

No two clusters should have the same name across datacenters in a given vCenter.

15.6 ESXi/vCenter System Requirements Edit source

For information about recommended hardware minimums, consult Book “Planning an Installation with Cloud Lifecycle Manager”, Chapter 3 “Recommended Hardware Minimums for the Example Configurations”, Section 3.2 “Recommended Hardware Minimums for an Entry-scale ESX KVM Model”.

15.7 Creating an ESX Cluster Edit source

Steps to create an ESX Cluster:

  1. Download the ESXi Hypervisor and vCenter Appliance from the VMware website.

  2. Install the ESXi Hypervisor.

  3. Configure the Management Interface.

  4. Enable the CLI and Shell access.

  5. Set the password and login credentials.

  6. Extract the vCenter Appliance files.

  7. The vCenter Appliance offers two ways to install the vCenter. The directory vcsa-ui-installer contains the graphical installer. The vcsa-cli-installer directory contains the command line installer. The remaining steps demonstrate using the vcsa-ui-installer installer.

  8. In the vcsa-ui-installer, click the installer to start installing the vCenter Appliance in the ESXi Hypervisor.

  9. Note the MANAGEMENT IP, USER ID, and PASSWORD of the ESXi Hypervisor.

  10. Assign an IP ADDRESS, USER ID, and PASSWORD to the vCenter server.

  11. Complete the installation.

  12. When the installation is finished, point your Web browser to the IP ADDRESS of the vCenter. Connect to the vCenter by clicking on link in the browser.

  13. Enter the information for the vCenter you just created: IP ADDRESS, USER ID, and PASSWORD.

  14. When connected, configure the following:

    • Datacenter

      1. Go to Home > Inventory > Hosts and Clusters

      2. Select File > New > Datacenter

      3. Rename the datacenter

    • Cluster

      1. Right-click a datacenter or directory in the vSphere Client and select New Cluster.

      2. Enter a name for the cluster.

      3. Choose cluster features.

    • Add a Host to Cluster

      1. In the vSphere Web Client, navigate to a datacenter, cluster, or directory within a datacenter.

      2. Right-click the datacenter, cluster, or directory and select Add Host.

      3. Type the IP address or the name of the host and click Next.

      4. Enter the administrator credentials and click Next.

      5. Review the host summary and click Next.

      6. Assign a license key to the host.

15.8 Configuring the Required Distributed vSwitches and Port Groups Edit source

The required Distributed vSwitches (DVS) and port groups can be created by using the vCenter graphical user interface (GUI) or by using the command line tool provided by python-networking-vsphere. The vCenter GUI is recommended.

OVSvApp virtual machines (VMs) give ESX installations the ability to leverage some of the advanced networking capabilities and other benefits OpenStack provides. In particular, OVSvApp allows for hosting VMs on ESX/ESXi hypervisors together with the flexibility of creating port groups dynamically on Distributed Virtual Switch.

A port group is a management object for aggregation of multiple ports (on a virtual switch) under a common configuration. A VMware port group is used to group together a list of ports in a virtual switch (DVS in this section) so that they can be configured all at once. The member ports of a port group inherit their configuration from the port group, allowing for configuration of a port by simply dropping it into a predefined port group.

The following sections cover configuring OVSvApp switches on ESX. More information about OVSvApp is available at https://wiki.openstack.org/wiki/Neutron/Networking-vSphere

The diagram below illustrates a typical configuration that uses OVSvApp and Distributed vSwitches.

Detailed instructions are shown in the following sections for four example installations and two command line procedures.

15.8.1 Creating ESXi TRUNK DVS and Required Portgroup Edit source

The process of creating an ESXi Trunk Distributed vSwitch (DVS) consists of three steps: create a switch, add host and physical adapters, and add a port group. Use the following detailed instructions to create a trunk DVS and a required portgroup. These instructions use a graphical user interface (GUI). The GUI menu options may vary slightly depending on the specific version of vSphere installed. Command line interface (CLI) instructions are below the GUI instructions.

15.8.1.1 Creating ESXi Trunk DVS with vSphere Web Client Edit source

  1. Create the switch.

    1. Using vSphere webclient, connect to the vCenter server.

    2. Under Hosts and cluster, right-click on the appropriate datacenter. Select Distributed Switch > New Distributed Switch.

    3. Name the switch TRUNK. Click Next.

    4. Select version 6.0.0 or larger. Click Next.

    5. Under Edit settings, lower the number of uplink ports to the lowest possible number (0 or 1). Uncheck Create a default port group. Click Next.

    6. Under Ready to complete, verify the settings are correct and click Finish.

  2. Add host and physical adapters.

    1. Under Networking find the DVS named TRUNK you just created. Right-click on it and select Manage hosts.

    2. Under Select task, select Add hosts. Click Next.

    3. Click New hosts.

    4. Select the CURRENT ESXI HOST and select OK. Click Next.

    5. Under Select network adapter tasks, select Manage advanced host settings and UNCHECK all other boxes. Click Next.

    6. Under Advanced host settings, check that the Maximum Number of Ports reads (auto). There is nothing else to do. Click Next.

    7. Under Ready to complete, verify that one and only one host is being added and click Finish.

  3. Add port group.

    1. Right-click on the TRUNK DVS that was just created (or modified) and select Distributed Port Group > New Distributed Port Group.

    2. Name the port group TRUNK-PG. Click Next.

    3. Under Configure settings select:

      • port binding > Static binding

      • port allocation > Elastic

      • vlan type > VLAN trunking with range of 1–4094.

    4. Check Customized default policies configuration. Click Next.

    5. Under Security use the following values:

      Setting

      Value

      promiscuous mode

      accept

      MAC address changes

      reject

      Forged transmits

      accept

    6. Set Autoexpand to true (port count growing).

    7. Skip Traffic shaping and click Next.

    8. Skip Teaming and fail over and click Next.

    9. Skip Monitoring and click Next.

    10. Under Miscellaneous there is nothing to be done. Click Next.

    11. Under Edit additional settings add a description if desired. Click Next.

    12. Under Ready to complete verify everything is as expected and click Finish.

15.8.2 Creating ESXi MGMT DVS and Required Portgroup Edit source

The process of creating an ESXi Mgmt Distributed vSwitch (DVS) consists of three steps: create a switch, add host and physical adapters, and add a port group. Use the following detailed instructions to create a mgmt DVS and a required portgroup.

  1. Create the switch.

    1. Using the vSphere webclient, connect to the vCenter server.

    2. Under Hosts and Cluster, right-click on the appropriate datacenter, and select Distributed Switch > New Distributed Switch

    3. Name the switch MGMT. Click Next.

    4. Select version 6.0.0 or higher. Click Next.

    5. Under Edit settings, select the appropriate number of uplinks. The MGMT DVS is what connects the ESXi host to the OpenStack management network. Uncheck Create a default port group. Click Next.

    6. Under Ready to complete, verify the settings are correct. Click Finish.

  2. Add host and physical adapters to Distributed Virtual Switch.

    1. Under Networking, find the MGMT DVS you just created. Right-click on it and select Manage hosts.

    2. Under Select task, select Add hosts. Click Next.

    3. Click New hosts.

    4. Select the current ESXi host and select OK. Click Next.

    5. Under Select network adapter tasks, select Manage physical adapters and UNCHECK all other boxes. Click Next.

    6. Under Manage physical network adapters, click on the interface you are using to connect the ESXi to the OpenStack management network. The name is of the form vmnic# (for example, vmnic0, vmnic1, etc.). When the interface is highlighted, select Assign uplink then select the uplink name to assign or auto assign. Repeat the process for each uplink physical NIC you will be using to connect to the OpenStack data network. Click Next.

    7. Verify that you understand and accept the impact shown by Analyze impact. Click Next.

    8. Verify that everything is correct and click on Finish.

  3. Add MGMT port group to switch.

    1. Right-click on the MGMT DVS and select Distributed Port Group > New Distributed Port Group.

    2. Name the port group MGMT-PG. Click Next.

    3. Under Configure settings, select:

      • port binding > Static binding

      • port allocation > Elastic

      • vlan type > None

      Click Next.

    4. Under Ready to complete, verify that everything is as expected and click Finish.

  4. Add GUEST port group to the switch.

    1. Right-click on the DVS (MGMT) that was just created (or modified). Select Distributed Port Group > New Distributed Port Group.

    2. Name the port group GUEST-PG. Click Next.

    3. Under Configure settings, select:

      • port binding > Static binding

      • port allocation > Elastic

      • vlan type > VLAN trunking The VLAN range corresponds to the VLAN ids being used by the OpenStack underlay. This is the same VLAN range as configured in the neutron.conf configuration file for the Neutron server.

    4. Select Customize default policies configuration. Click Next.

    5. Under Security, use the following settings:

      setting

      value

      promiscuous mode

      accept

      MAC address changes

      reject

      Forged transmits

      accept

    6. Skip Traffic shaping and click Next.

    7. Under Teaming and fail over, make changes appropriate for your network and deployment.

    8. Skip Monitoring and click Next.

    9. Skip Miscellaneous and click Next.

    10. Under Edit addition settings, add a description if desired. Click Next.

    11. Under Ready to complete, verify everything is as expected. Click Finish.

  5. Add ESX-CONF port group.

    1. Right-click on the DVS (MGMT) that was just created (or modified). Select Distributed Port Group > New Distributed Port Group.

    2. Name the port group ESX-CONF-PG. Click Next.

    3. Under Configure settings, select:

      • port binding > Static binding

      • port allocation > Elastic

      • vlan type > None

      Click Next.

      • port binding > Static binding

      • port allocation > Elastic

      • vlan type > None

      Click Next.

    4. Under Ready to complete, verify that everything is as expected and click Finish.

15.8.3 Configuring OVSvApp Network Resources Using Ansible-Playbook Edit source

The Ardana ansible playbook neutron-create-ovsvapp-resources.yml can be used to create Distributed Virtual Switches and Port Groups on a vCenter cluster.

The playbook requires the following inputs:

  • vcenter_username

  • vcenter_encrypted_password

  • vcenter_ip

  • vcenter_port (default 443)

  • vc_net_resources_location This is the path to a file which contains the definition of the resources to be created. The definition is in JSON format.

In order to execute the playbook from the Cloud Lifecycle Manager, the python-networking-vsphere package must be installed.

tux > sudo zypper install python-networking-vsphere

Running the playbook:

ardana > ansible-playbook neutron-create-ovsvapp-resources.yml \
-i hosts/verb_hosts -vvv -e 'variable_host=localhost
vcenter_username=USERNAME
vcenter_encrypted_password=ENCRYPTED_PASSWORD
vcenter_ip=IP_ADDRESS
vcenter_port=443
vc_net_resources_location=LOCATION_TO_RESOURCE_DEFINITION_FILE
'

The RESOURCE_DEFINITION_FILE is in JSON format and contains the resources to be created.

Sample file contents:

{
  "datacenter_name": "DC1",
  "host_names": [
    "192.168.100.21",
    "192.168.100.222"
  ],
  "network_properties": {
    "switches": [
      {
        "type": "dvSwitch",
        "name": "TRUNK",
        "pnic_devices": [],
        "max_mtu": "1500",
        "description": "TRUNK DVS for ovsvapp.",
        "max_ports": 30000
      },
      {
        "type": "dvSwitch",
        "name": "MGMT",
        "pnic_devices": [
          "vmnic1"
        ],
        "max_mtu": "1500",
        "description": "MGMT DVS for ovsvapp. Uses 'vmnic0' to connect to OpenStack Management network",
        "max_ports": 30000
      }
    ],
    "portGroups": [
      {
        "name": "TRUNK-PG",
        "vlan_type": "trunk",
        "vlan_range_start": "1",
        "vlan_range_end": "4094",
        "dvs_name": "TRUNK",
        "nic_teaming": null,
        "allow_promiscuous": true,
        "forged_transmits": true,
        "auto_expand": true,
        "description": "TRUNK port group. Configure as trunk for vlans 1-4094. Default nic_teaming selected."
      },
      {
        "name": "MGMT-PG",
        "dvs_name": "MGMT",
        "nic_teaming": null,
        "description": "MGMT port group. Configured as type 'access' (vlan with vlan_id = 0, default). Default nic_teaming. Promiscuous false, forged_transmits default"
      },
      {
        "name": "GUEST-PG",
        "dvs_name": "GUEST",
        "vlan_type": "MGMT",
        "vlan_range_start": 100,
        "vlan_range_end": 200,
        "nic_teaming": null,
        "allow_promiscuous": true,
        "forged_transmits": true,
        "auto_expand": true,
        "description": "GUEST port group. Configure for vlans 100 through 200."
      },
      {
        "name": "ESX-CONF-PG",
        "dvs_name": "MGMT",
        "nic_teaming": null,
        "description": "ESX-CONF port group. Configured as type 'access' (vlan with vlan_id = 0, default)."
      }
    ]
  }
}

15.8.4 Configuring OVSVAPP Using Python-Networking-vSphere Edit source

Scripts can be used with the Networking-vSphere Project. The scripts automate some of the process of configuring OVSvAPP from the command line. The following are help entries for two of the scripts:

tux > cd /opt/repos/networking-vsphere
 tux > ovsvapp-manage-dvs -h
usage: ovsvapp-manage-dvs [-h] [--tcp tcp_port]
                          [--pnic_devices pnic_devices [pnic_devices ...]]
                          [--max_mtu max_mtu]
                          [--host_names host_names [host_names ...]]
                          [--description description] [--max_ports max_ports]
                          [--cluster_name cluster_name] [--create]
                          [--display_spec] [-v]
                          dvs_name vcenter_user vcenter_password vcenter_ip
                          datacenter_name
positional arguments:
  dvs_name              Name to use for creating the DVS
  vcenter_user          Username to be used for connecting to vCenter
  vcenter_password      Password to be used for connecting to vCenter
  vcenter_ip            IP address to be used for connecting to vCenter
  datacenter_name       Name of data center where the DVS will be created
optional arguments:
  -h, --help            show this help message and exit
  --tcp tcp_port        TCP port to be used for connecting to vCenter
  --pnic_devices pnic_devices [pnic_devices ...]
                        Space separated list of PNIC devices for DVS
  --max_mtu max_mtu     MTU to be used by the DVS
  --host_names host_names [host_names ...]
                        Space separated list of ESX hosts to add to DVS
  --description description
                        DVS description
  --max_ports max_ports
                        Maximum number of ports allowed on DVS
  --cluster_name cluster_name
                        Cluster name to use for DVS
  --create              Create DVS on vCenter
  --display_spec        Print create spec of DVS
 -v                    Verbose output
tux > cd /opt/repos/networking-vsphere
 tux > ovsvapp-manage-dvpg -h
usage: ovsvapp-manage-dvpg [-h] [--tcp tcp_port] [--vlan_type vlan_type]
                           [--vlan_id vlan_id]
                           [--vlan_range_start vlan_range_start]
                           [--vlan_range_stop vlan_range_stop]
                           [--description description] [--allow_promiscuous]
                           [--allow_forged_transmits] [--notify_switches]
                           [--network_failover_detection]
                           [--load_balancing {loadbalance_srcid,loadbalance_ip,loadbalance_srcmac,loadbalance_loadbased,failover_explicit}]
                           [--create] [--display_spec]
                           [--active_nics ACTIVE_NICS [ACTIVE_NICS ...]] [-v]
                           dvpg_name vcenter_user vcenter_password vcenter_ip
                           dvs_name
positional arguments:
  dvpg_name             Name to use for creating theDistributed Virtual Port
                        Group (DVPG)
  vcenter_user          Username to be used for connecting to vCenter
  vcenter_password      Password to be used for connecting to vCenter
  vcenter_ip            IP address to be used for connecting to vCenter
  dvs_name              Name of the Distributed Virtual Switch (DVS) to create
                        the DVPG in
optional arguments:
  -h, --help            show this help message and exit
  --tcp tcp_port        TCP port to be used for connecting to vCenter
  --vlan_type vlan_type
                        Vlan type to use for the DVPG
  --vlan_id vlan_id     Vlan id to use for vlan_type='vlan'
  --vlan_range_start vlan_range_start
                        Start of vlan id range for vlan_type='trunk'
  --vlan_range_stop vlan_range_stop
                        End of vlan id range for vlan_type='trunk'
  --description description
                        DVPG description
  --allow_promiscuous   Sets promiscuous mode of DVPG
  --allow_forged_transmits
                        Sets forge transmit mode of DVPG
  --notify_switches     Set nic teaming 'notify switches' to True.
  --network_failover_detection
                        Set nic teaming 'network failover detection' to True
  --load_balancing {loadbalance_srcid,loadbalance_ip,loadbalance_srcmac,loadbalance_loadbased,failover_explicit}
                        Set nic teaming load balancing algorithm.
                        Default=loadbalance_srcid
  --create              Create DVPG on vCenter
  --display_spec        Send DVPG's create spec to OUTPUT
  --active_nics ACTIVE_NICS [ACTIVE_NICS ...]
                        Space separated list of active nics to use in DVPG nic
                        teaming
 -v                    Verbose output

15.9 Create a SUSE-based Virtual Appliance Template in vCenter Edit source

  1. Download the SLES12-SP3 ISO image (SLE-12-SP4-Server-DVD-x86_64-GM-DVD1.iso) from https://www.suse.com/products/server/download/. You need to sign in or create a SUSE customer service account before downloading.

  2. Create a new Virtual Machine in vCenter Resource Pool.

  3. Configure the Storage selection.

  4. Configure the Guest Operating System.

  5. Create a Disk.

  6. Ready to Complete.

  7. Edit Settings before booting the VM with additional Memory, typically 16GB or 32GB, though large scale environments may require larger memory allocations.

  8. Edit Settings before booting the VM with additional Network Settings. Ensure there are four network adapters, one each for TRUNK, MGMT, ESX-CONF, and GUEST.

  9. Attach the ISO image to the DataStore.

  10. Configure the 'disk.enableUUID=TRUE' flag in the General - Advanced Settings.

  11. After attaching the CD/DVD drive with the ISO image and completing the initial VM configuration, power on the VM by clicking the Play button on the VM's summary page.

  12. Click Installation when the VM boots from the console window.

  13. Accept the License agreement, language and Keyboard selection.

  14. Select the System Role to Xen Virtualization Host.

  15. Select the 'Proposed Partitions' in the Suggested Partition screen.

  16. Edit the Partitions to select the 'LVM' Mode and then select the 'ext4' filesystem type.

  17. Increase the size of the root partition from 10GB to 60GB.

  18. Create an additional logical volume to accommodate the LV_CRASH volume (15GB). Do not mount the volume at this time, it will be used later.

  19. Configure the Admin User/Password and User name.

  20. Installation Settings (Disable Firewall and enable SSH).

  21. The operating system will be successfully installed and the VM will reboot.

  22. Check that the contents of the ISO files are copied to the locations shown below on your Cloud Lifecycle Manager. This may already be completed on the Cloud Lifecycle Manager.

    • The contents of the SLES SDK ISO (SLE-12-SP3-SDK-DVD-x86_64-GM-DVD1.iso) must be mounted or copied to /opt/ardana_packager/ardana/sles12/zypper/SDK/ (create the directory if it is missing). If you choose to mount the ISO, we recommend creating an /etc/fstab entry to ensure the ISO is mounted after a reboot.

    • Mount or copy the contents of SLE-12-SP3-Server-DVD-x86_64-GM-DVD1.iso to /opt/ardana_packager/ardana/sles12/zypper/OS/ (create the directory if it is missing).

    • Mount or copy the contents of SLE-12-SP3-SDK-DVD-x86_64-GM-DVD1.iso to /opt/ardana_packager/ardana/sles12/zypper/SDK/.

  23. Log in to the VM with the configured user credentials.

  24. The VM must be set up before a template can be created with it. The IP addresses configured here are temporary and will need to be reconfigured as VMs are created using this template. The temporary IP address should not overlap with the network range for the MGMT network.

    1. The VM should now have four network interfaces. Configure them as follows:

      1. ardana > cd /etc/sysconfig/network
               tux > sudo ls

        The directory will contain the files: ifcfg-br0, ifcfg-br1, ifcfg-br2, ifcfg-br3, ifcfg-eth0, ifcfg-eth1, ifcfg-eth2, and ifcfg-eth3.

      2. If you have configured a default route while installing the VM, then there will be a routes file.

      3. Note the IP addresses configured for MGMT.

      4. Configure the temporary IP for the MGMT network. Edit the ifcfg-eth1 file.

        1. tux > sudo vi /etc/sysconfig/network/ifcfg-eth1
          BOOTPROTO='static'
          BROADCAST=''
          ETHTOOL_OPTIONS=''
          IPADDR='192.168.24.132/24' (Configure the IP address of the MGMT Interface)
          MTU=''
          NETWORK=''
          REMOTE_IPADDR=''
          STARTMODE='auto'
      5. Edit the ifcfg-eth0, ifcfg-eth2, and ifcfg-eth3 files.

        1. tux > sudo vi /etc/sysconfig/network/ifcfg-eth0
          BOOTPROTO='static'
          BROADCAST=''
          ETHTOOL_OPTIONS=''
          IPADDR=''
          MTU=''
          NETWORK=''
          REMOTE_IPADDR=''
          STARTMODE='auto'
        2. tux > sudo vi /etc/sysconfig/network/ifcfg-eth2
          BOOTPROTO=''
          BROADCAST=''
          ETHTOOL_OPTIONS=''
          IPADDR=''
          MTU=''
          NAME='VMXNET3 Ethernet Controller'
          NETWORK=''
          REMOTE_IPADDR=''
          STARTMODE='auto'
        3. tux > sudo vi /etc/sysconfig/network/ifcfg-eth3
          BOOTPROTO=''
          BROADCAST=''
          ETHTOOL_OPTIONS=''
          IPADDR=''
          MTU=''
          NAME='VMXNET3 Ethernet Controller'
          NETWORK=''
          REMOTE_IPADDR=''
          STARTMODE='auto'
        4. If the default route is not configured, add a default route file manually.

          1. Create a file routes in /etc/sysconfig/network.

          2. Edit the file to add your default route.

            tux > sudo sudo vi routes
                       default 192.168.24.140 - -
      6. Delete all the bridge configuration files, which are not required: ifcfg-br0, ifcfg-br1, ifcfg-br2, and ifcfg-br3.

    2. Add ardana user and home directory if that is not your default user. The username and password should be ardana/ardana.

      tux > sudo useradd -m ardana
            tux > sudo passwd ardana
    3. Create a ardana usergroup in the VM if it does not exist.

      1. Check for an existing ardana group.

        tux > sudo groups ardana
      2. Add ardana group if necessary.

        tux > sudo groupadd ardana
      3. Add ardana user to the ardana group.

        tux > sudo gpasswd -a ardana ardana
    4. Allow the ardana user to sudo without password. Setting up sudo on SLES is covered in the SUSE documentation at https://documentation.suse.com/sles/12-SP5/single-html/SLES-admin/#sec-sudo-conf. We recommend creating user specific sudo config files in the /etc/sudoers.d directory. Create an /etc/sudoers.d/ardana config file with the following content to allow sudo commands without the requirement of a password.

      ardana ALL=(ALL) NOPASSWD:ALL
    5. Add the Zypper repositories using the ISO-based repositories created previously. Change the value of deployer_ip if necessary.

      tux > sudo DEPLOYER_IP=192.168.24.140
           tux > sudo zypper addrepo --no-gpgcheck --refresh \
           http://$deployer_ip:79/ardana/sles12/zypper/OS SLES-OS
           tux > sudo zypper addrepo --no-gpgcheck --refresh \
           http://$deployer_ip:79/ardana/sles12/zypper/SDK SLES-SDK

      Verify that the repositories have been added.

      tux > zypper repos --detail
    6. Set up SSH access that does not require a password to the temporary IP address that was configured for eth1 .

      When you have started the installation using the Cloud Lifecycle Manager or if you are adding a SLES node to an existing cloud, the Cloud Lifecycle Manager public key needs to be copied to the SLES node. You can do this by copying ~/.ssh/authorized_keys from another node in the cloud to the same location on the SLES node. If you are installing a new cloud, this file will be available on the nodes after running the bm-reimage.yml playbook.

      Important
      Important

      Ensure that there is global read access to the file ~/.ssh/authorized_keys.

      Test passwordless ssh from the Cloud Lifecycle Manager and check your ability to remotely execute sudo commands.

      tux > ssh ardana@IP_OF_SLES_NODE_eth1
           "sudo tail -5 /var/log/messages"
  25. Shutdown the VM and create a template out of the VM appliance for future use.

  26. The VM Template will be saved in your vCenter Datacenter and you can view it from VMS and Templates menu. Note that menu options will vary slightly depending on the version of vSphere that is deployed.

15.10 ESX Network Model Requirements Edit source

For this model the following networks are needed:

  • MANAGEMENT-NET : This is an untagged network this is used for the control plane as well as the esx-compute proxy and ovsvapp VMware instance. It is tied to the MGMT DVS/PG in vSphere.

  • EXTERNAL-API_NET : This is a tagged network for the external/public API. There is no difference in this model from those without ESX and there is no additional setup needed in vSphere for this network.

  • EXTERNAL-VM-NET : This is a tagged network used for Floating IP (FIP) assignment to running instances. There is no difference in this model from those without ESX and there is no additional setup needed in vSphere for this network.

  • GUEST-NET : This is a tagged network used internally for neutron. It is tied to the GUEST PG in vSphere.

  • ESX-CONF-NET : This is a separate configuration network for ESX that must be reachable via the MANAGEMENT-NET. It is tied to the ESX-CONF PG in vSphere.

  • TRUNK-NET : This is an untagged network used internally for ESX. It is tied to the TRUNC DVS/PG in vSphere.

15.11 Creating and Configuring Virtual Machines Based on Virtual Appliance Template Edit source

The following process for creating and configuring VMs from the vApp template should be repeated for every cluster in the DataCenter. Each cluster should host a Nova Proxy VM, and each host in a cluster should have an OVSvApp VM running. The following method uses the vSphere Client Management Tool to deploy saved templates from the vCenter.

  1. Identify the cluster that you want Nova Proxy to manage.

  2. Create a VM from the template on a chosen cluster.

  3. The first VM that was deployed will be the Nova Compute Proxy VM. This VM can reside on any HOST inside a cluster. There should be only one instance of this VM in a cluster.

  4. The Nova Compute Proxy will use only two of the four interfaces configured previously (ESX_CONF and MANAGEMENT).

    Note
    Note

    Do not swap the interfaces. They must be in the specified order (ESX_CONF is eth0, MGMT is eth1).

  5. After the VM has been deployed, log in to it with ardana/ardana credentials. Log in to the VM with SSH using the MGMT IP address. Make sure that all root level commands work with sudo. This is required for the Cloud Lifecycle Manager to configure the appliance for services and networking.

  6. Install another VM from the template and name it OVSvApp-VM1-HOST1. (You can add a suffix with the host name to identify the host it is associated with).

    Note
    Note

    The VM must have four interfaces configured in the right order. The VM must be accessible from the Management Network through SSH from the Cloud Lifecycle Manager.

    • /etc/sysconfig/network/ifcfg-eth0 is ESX_CONF.

    • /etc/sysconfig/network/ifcfg-eth1 is MGMT.

    • /etc/sysconfig/network/ifcfg-eth2 is TRUNK.

    • /etc/sysconfig/network/ifcfg-eth3 is GUEST.

  7. If there is more than one HOST in the cluster, deploy another VM from the Template and name it OVSvApp-VM2-HOST2.

  8. If the OVSvApp VMs end up on the same HOST, then manually separate the VMs and follow the instructions below to add rules for High Availability (HA) and Distributed Resource Scheduler (DRS).

    Note
    Note

    HA seeks to minimize system downtime and data loss. See also Book “Planning an Installation with Cloud Lifecycle Manager”, Chapter 4 “High Availability”. DRS is a utility that balances computing workloads with available resources in a virtualized environment.

  9. When installed from a template to a cluster, the VM will not be bound to a particular host if you have more than one Hypervisor. The requirement for the OVSvApp is that there be only one OVSvApp Appliance per host and that it should be constantly bound to the same host. DRS or VMotion should not be allowed to migrate the VMs from the existing HOST. This would cause major network interruption. In order to achieve this we need to configure rules in the cluster HA and DRS settings.

    Note
    Note

    VMotion enables the live migration of running virtual machines from one physical server to another with zero downtime, continuous service availability, and complete transaction integrity.

  10. Configure rules for OVSvApp VMs.

    1. Configure vSphere HA - Virtual Machine Options.

    2. Use Cluster Setting must be disabled.

    3. VM should be Power-On.

  11. Configure Cluster DRS Groups/Rules.

    1. Configure vSphere DRS - DRS Group Manager.

    2. Create a DRS Group for the OVSvApp VMs.

    3. Add VMs to the DRS Group.

    4. Add appropriate Rules to the DRS Groups.

  12. All three VMs are up and running. Following the preceding process, there is one Nova Compute Proxy VM per cluster, and OVSvAppVM1 and OVSvAppVM2 on each HOST in the cluster.

  13. Record the configuration attributes of the VMs.

    • Nova Compute Proxy VM:

      • Cluster Name where this VM is located

      • Management IP Address

      • VM Name The actual name given to the VM to identify it.

    • OVSvAppVM1

      • Cluster Name where this VM is located

      • Management IP Address

      • esx_hostname that this OVSvApp is bound to

      • cluster_dvs_mapping The Distributed vSwitch name created in the datacenter for this particular cluster.

        Example format:

        DATA_CENTER/host/CLUSTERNAME: DVS-NAME Do not substitute for host'. It is a constant.

    • OVSvAppVM2:

      • Cluster Name where this VM is located

      • Management IP Address

      • esx_hostname that this OVSvApp is bound to

      • cluster_dvs_mapping The Distributed vSwitch name created in the datacenter for this particular cluster.

        Example format:

        DATA_CENTER/host/CLUSTERNAME: DVS-NAME Do not substitute for host'. It is a constant.

15.12 Collect vCenter Credentials and UUID Edit source

  • Obtain the vCenter UUID from vSphere with the URL shown below:

    https://VCENTER-IP/mob/?moid=ServiceInstance&doPath=content.about

    Select the field instanceUUID. Copy and paste the value of # field instanceUUID.

  • Record the UUID

  • Record the vCenter Password

  • Record the vCenter Management IP

  • Record the DataCenter Name

  • Record the Cluster Name

  • Record the DVS (Distributed vSwitch) Name

15.13 Edit Input Models to Add and Configure Virtual Appliances Edit source

The following steps should be used to edit the Ardana input model data to add and configure the Virtual Appliances that were just created. The process assumes that the SUSE OpenStack Cloud is deployed and a valid Cloud Lifecycle Manager is in place.

  1. Edit the following files in ~/openstack/my_cloud/definition/data/: servers.yml, disks_app_vm.yml, and pass_through.yml. Fill in attribute values recorded in the previous step.

  2. Follow the instructions in pass_through.yml to encrypt your vCenter password using an encryption key.

  3. Export an environment variable for the encryption key.

    ARDANA_USER_PASSWORD_ENCRYPT_KEY=ENCRYPTION_KEY
  4. Run ~ardana/openstack/ardana/ansible/ardanaencrypt.py script. It will prompt for unencrypted value?. Enter the unencrypted vCenter password and it will return an encrypted string.

  5. Copy and paste the encrypted password string in the pass_through.yml file as a value for the password field enclosed in double quotes.

  6. Enter the username, ip, and id of the vCenter server in the Global section of the pass_through.yml file. Use the values recorded in the previous step.

  7. In the servers section of the pass_through.yml file, add the details about the Nova Compute Proxy and OVSvApp VMs that was recorded in the previous step.

    # Here the 'id' refers to the name of the node running the
          # esx-compute-proxy. This is identical to the 'servers.id' in
          # servers.yml.
          # NOTE: There should be one esx-compute-proxy node per ESX
          # resource pool or cluster.
          # cluster_dvs_mapping in the format
          # 'Datacenter-name/host/Cluster-Name:Trunk-DVS-Name'
          # Here 'host' is a string and should not be changed or
          # substituted.
          # vcenter_id is same as the 'vcenter-uuid' obtained in the global
          # section.
          # 'id': is the name of the appliance manually installed
          # 'vcenter_cluster': Name of the vcenter target cluster
          # esx_hostname: Name of the esx host hosting the ovsvapp
          # NOTE: For every esx host in a cluster there should be an ovsvapp
          # instance running.
          id: esx-compute1
          data:
            vmware:
              vcenter_cluster: <vmware cluster1 name>
              vcenter_id: <vcenter-uuid>
        -
          id: ovsvapp1
          data:
            vmware:
              vcenter_cluster: <vmware cluster1 name>
              cluster_dvs_mapping: <cluster dvs mapping>
              esx_hostname: <esx hostname hosting the ovsvapp>
              vcenter_id: <vcenter-uuid>
        -
          id: ovsvapp2
          data:
            vmware:
              vcenter_cluster: <vmware cluster1 name>
              cluster_dvs_mapping: <cluster dvs mapping>
              esx_hostname: <esx hostname hosting the ovsvapp>
              vcenter_id: <vcenter-uuid>

    The VM id string should match exactly with the data written in the servers.yml file.

  8. Edit the servers.yml file, adding the Nova Proxy VM and OVSvApp information recorded in the previous step.

    # Below entries shall be added by the user
        # for entry-scale-kvm-esx after following
        # the doc instructions in creating the
        # esx-compute-proxy VM Appliance and the
        # esx-ovsvapp VM Appliance.
        # Added just for the reference
        # NOTE: There should be one esx-compute per
        # Cluster and one ovsvapp per Hypervisor in
        # the Cluster.
        # id - is the name of the virtual appliance
        # ip-addr - is the Mgmt ip address of the appliance
        # The values shown below are examples and has to be
        # substituted based on your setup.
        # Nova Compute proxy node
        - id: esx-compute1
          server-group: RACK1
          ip-addr: 192.168.24.129
          role: ESX-COMPUTE-ROLE
        # OVSVAPP node
        - id: ovsvapp1
          server-group: RACK1
          ip-addr: 192.168.24.130
          role: OVSVAPP-ROLE
        - id: ovsvapp2
          server-group: RACK1
          ip-addr: 192.168.24.131
          role: OVSVAPP-ROLE

    Examples of pass_through.yml and servers.yml files:

    pass_through.yml
    product:
      version: 2
    pass-through:
      global:
        vmware:
          - username: administrator@vsphere.local
            ip: 10.84.79.3
            port: '443'
            cert_check: false
            password: @hos@U2FsdGVkX19aqGOUYGgcAIMQSN2lZ1X+gyNoytAGCTI=
            id: a0742a39-860f-4177-9f38-e8db82ad59c6
      servers:
        - data:
            vmware:
              vcenter_cluster: QE
              vcenter_id: a0742a39-860f-4177-9f38-e8db82ad59c6
          id: lvm-nova-compute1-esx01-qe
        - data:
            vmware:
              vcenter_cluster: QE
              cluster_dvs_mapping: 'PROVO/host/QE:TRUNK-DVS-QE'
              esx_hostname: esx01.qe.provo
              vcenter_id: a0742a39-860f-4177-9f38-e8db82ad59c6
          id: lvm-ovsvapp1-esx01-qe
        - data:
            vmware:
              vcenter_cluster: QE
              cluster_dvs_mapping: 'PROVO/host/QE:TRUNK-DVS-QE'
              esx_hostname: esx02.qe.provo
              vcenter_id: a0742a39-860f-4177-9f38-e8db82ad59c6
              id: lvm-ovsvapp2-esx02-qe
    servers.yml
    product:
      version: 2
    servers:
      - id: deployer
        ilo-ip: 192.168.10.129
        ilo-password: 8hAcPMne
        ilo-user: CLM004
        ip-addr: 192.168.24.125
        is-deployer: true
        mac-addr: '8c:dc:d4:b4:c5:4c'
        nic-mapping: MY-2PORT-SERVER
        role: DEPLOYER-ROLE
        server-group: RACK1
      - id: controller3
        ilo-ip: 192.168.11.52
        ilo-password: 8hAcPMne
        ilo-user: HLM004
        ip-addr: 192.168.24.128
        mac-addr: '8c:dc:d4:b5:ed:b8'
        nic-mapping: MY-2PORT-SERVER
        role: CONTROLLER-ROLE
        server-group: RACK1
      - id: controller2
        ilo-ip: 192.168.10.204
        ilo-password: 8hAcPMne
        ilo-user: HLM004
        ip-addr: 192.168.24.127
        mac-addr: '8c:dc:d4:b5:ca:c8'
        nic-mapping: MY-2PORT-SERVER
        role: CONTROLLER-ROLE
        server-group: RACK2
      - id: controller1
        ilo-ip: 192.168.11.57
        ilo-password: 8hAcPMne
        ilo-user: CLM004
        ip-addr: 192.168.24.126
        mac-addr: '5c:b9:01:89:c6:d8'
        nic-mapping: MY-2PORT-SERVER
        role: CONTROLLER-ROLE
        server-group: RACK3
      # Nova compute proxy for QE cluster added manually
      - id: lvm-nova-compute1-esx01-qe
        server-group: RACK1
        ip-addr: 192.168.24.129
        role: ESX-COMPUTE-ROLE
      # OVSvApp VM for QE cluster added manually
      # First ovsvapp vm in esx01 node
      - id: lvm-ovsvapp1-esx01-qe
        server-group: RACK1
        ip-addr: 192.168.24.132
        role: OVSVAPP-ROLE
      # Second ovsvapp vm in esx02 node
      - id: lvm-ovsvapp2-esx02-qe
        server-group: RACK1
        ip-addr: 192.168.24.131
        role: OVSVAPP-ROLE
    baremetal:
      subnet: 192.168.24.0
      netmask: 255.255.255.0
  9. Edit the disks_app_vm.yml file based on your lvm configuration. The attributes of Volume Group, Physical Volume, and Logical Volumes must be edited based on the LVM configuration of the VM.

    When you partitioned LVM during installation, you received Volume Group name, Physical Volume name and Logical Volumes with their partition sizes.

    This information can be retrieved from any of the VMs (Nova Proxy VM or the OVSvApp VM):

    tux > sudo pvdisplay
    # — Physical volume —
        # PV Name /dev/sda1
        # VG Name system
        # PV Size 80.00 GiB / not usable 3.00 MiB
        # Allocatable yes
        # PE Size 4.00 MiB
        # Total PE 20479
        # Free PE 511
        # Allocated PE 19968
        # PV UUID 7Xn7sm-FdB4-REev-63Z3-uNdM-TF3H-S3ZrIZ

    The Physical Volume Name is /dev/sda1. And the Volume Group Name is system.

    To find Logical Volumes:

    tux > sudo fdisk -l
    # Disk /dev/sda: 80 GiB, 85899345920 bytes, 167772160 sectors
        # Units: sectors of 1 * 512 = 512 bytes
        # Sector size (logical/physical): 512 bytes / 512 bytes
        # I/O size (minimum/optimal): 512 bytes / 512 bytes
        # Disklabel type: dos
        # Disk identifier: 0x0002dc70
        # Device Boot Start End Sectors Size Id Type
        # /dev/sda1 * 2048 167772159 167770112 80G 8e Linux LVM
        # Disk /dev/mapper/system-root: 60 GiB, 64424509440 bytes,
        # 125829120 sectors
        # Units: sectors of 1 * 512 = 512 bytes
        # Sector size (logical/physical): 512 bytes / 512 bytes
        # I/O size (minimum/optimal): 512 bytes / 512 bytes
        # Disk /dev/mapper/system-swap: 2 GiB, 2147483648 bytes, 4194304 sectors
        # Units: sectors of 1 * 512 = 512 bytes
        # Sector size (logical/physical): 512 bytes / 512 bytes
        # I/O size (minimum/optimal): 512 bytes / 512 bytes
        # Disk /dev/mapper/system-LV_CRASH: 16 GiB, 17179869184 bytes,
        # 33554432 sectors
        # Units: sectors of 1 * 512 = 512 bytes
        # Sector size (logical/physical): 512 bytes / 512 bytes
        # I/O size (minimum/optimal): 512 bytes / 512 bytes
        # NOTE: Even though we have configured the SWAP partition, it is
        # not required to be configured in here. Just configure the root
        # and the LV_CRASH partition
    • The line with /dev/mapper/system-root: 60 GiB, 64424509440 bytes indicates that the first logical partition is root.

    • The line with /dev/mapper/system-LV_CRASH: 16 GiB, 17179869184 bytes indicates that the second logical partition is LV_CRASH.

    • The line with /dev/mapper/system-swap: 2 GiB, 2147483648 bytes, 4194304 sectors indicates that the third logical partition is swap.

  10. Edit the disks_app_vm.yml file. It is not necessary to configure the swap partition.

    volume-groups:
        - name: system (Volume Group Name)
          physical-volumes:
           - /dev/sda1 (Physical Volume Name)
          logical-volumes:
            - name: root   ( Logical Volume 1)
              size: 75%    (Size in percentage)
              fstype: ext4 ( filesystem type)
              mount: /     ( Mount point)
            - name: LV_CRASH   (Logical Volume 2)
              size: 20%        (Size in percentage)
              mount: /var/crash (Mount point)
              fstype: ext4      (filesystem type)
              mkfs-opts: -O large_file

    An example disks_app_vm.yml file:

    disks_app_vm.yml
    ---
      product:
        version: 2
      disk-models:
      - name: APP-VM-DISKS
        # Disk model to be used for application vms such as nova-proxy and ovsvapp
        # /dev/sda1 is used as a volume group for /, /var/log and /var/crash
        # Additional disks can be added to either volume group
        #
        # NOTE: This is just an example file and has to filled in by the user
        # based on the lvm partition map for their virtual appliance
        # While installing the operating system opt for the LVM partition and
        # create three partitions as shown below
        # Here is an example partition map
        # In this example we have three logical partitions
        # root partition (75%)
        # swap (5%) and
        # LV_CRASH (20%)
        # Run this command 'sudo pvdisplay' on the virtual appliance to see the
        # output as shown below
        #
        # — Physical volume —
        # PV Name /dev/sda1
        # VG Name system
        # PV Size 80.00 GiB / not usable 3.00 MiB
        # Allocatable yes
        # PE Size 4.00 MiB
        # Total PE 20479
        # Free PE 511
        # Allocated PE 19968
        # PV UUID 7Xn7sm-FdB4-REev-63Z3-uNdM-TF3H-S3ZrIZ
        #
        # Next run the following command on the virtual appliance
        #
        # sudo fdisk -l
        # The output will be as shown below
        #
        # Disk /dev/sda: 80 GiB, 85899345920 bytes, 167772160 sectors
        # Units: sectors of 1 * 512 = 512 bytes
        # Sector size (logical/physical): 512 bytes / 512 bytes
        # I/O size (minimum/optimal): 512 bytes / 512 bytes
        # Disklabel type: dos
        # Disk identifier: 0x0002dc70
        # Device Boot Start End Sectors Size Id Type
        # /dev/sda1 * 2048 167772159 167770112 80G 8e Linux LVM
        # Disk /dev/mapper/system-root: 60 GiB, 64424509440 bytes,
        # 125829120 sectors
        # Units: sectors of 1 * 512 = 512 bytes
        # Sector size (logical/physical): 512 bytes / 512 bytes
        # I/O size (minimum/optimal): 512 bytes / 512 bytes
        # Disk /dev/mapper/system-swap: 2 GiB, 2147483648 bytes, 4194304 sectors
        # Units: sectors of 1 * 512 = 512 bytes
        # Sector size (logical/physical): 512 bytes / 512 bytes
        # I/O size (minimum/optimal): 512 bytes / 512 bytes
        # Disk /dev/mapper/system-LV_CRASH: 16 GiB, 17179869184 bytes,
        # 33554432 sectors
        # Units: sectors of 1 * 512 = 512 bytes
        # Sector size (logical/physical): 512 bytes / 512 bytes
        # I/O size (minimum/optimal): 512 bytes / 512 bytes
        # NOTE: Even though we have configured the SWAP partition, it is
        # not required to be configured in here. Just configure the root
        # and the LV_CRASH partition
        volume-groups:
          - name: system
            physical-volumes:
             - /dev/sda1
            logical-volumes:
              - name: root
                size: 75%
                fstype: ext4
                mount: /
              - name: LV_CRASH
                size: 20%
                mount: /var/crash
                fstype: ext4
                mkfs-opts: -O large_file

15.14 Running the Configuration Processor With Applied Changes Edit source

If the changes are being applied to a previously deployed cloud, then after the previous section is completed, the Configuration Processor should be run with the changes that were applied.

  1. Run the Configuration Processor

    ardana > cd ~/ardana/ansible
    ardana > ansible-playbook -i hosts/localhost config-processor-run.yml \
    -e remove_deleted_servers="y" -e free_unused_addresses="y"
  2. ardana > ansible-playbook -i hosts/localhost ready-deployment.yml
  3. Run the site.yml playbook against only the VMs that were added.

    ardana > cd ~/scratch/ansible/next/ardana/ansible/
    ardana > ansible-playbook -i hosts/verb_hosts site.yml --extra-vars \
    "hux_svc_ignore_stop":true --limit hlm004-cp1-esx-comp0001-mgmt, \
    hlm004-cp1-esx-ovsvapp0001-mgmt,hlm004-cp1-esx-ovsvapp0002-mgmt

If the changes are being applied ahead of deploying a new (greenfield) cloud, then after the previous section is completed, the following steps should be run.

  1. Run the Configuration Processor

    ardana > cd ~/ardana/ansible
    	ardana > ansible-playbook -i hosts/localhost config-processor-run.yml
  2. ardana > ansible-playbook -i hosts/localhost ready-deployment.yml
  3. Run the site.yml playbook against only the VMs that were added.

    ardana > cd ~/scratch/ansible/next/ardana/ansible/
    	ardana > ansible-playbook -i hosts/verb_hosts site.yml

15.15 Test the ESX-OVSvApp Environment Edit source

When all of the preceding installation steps have been completed, test the ESX-OVSvApp environment with the following steps:

  1. SSH to the Controller

  2. Source the service.osrc file

  3. Create a Network

  4. Create a Subnet

  5. Create a VMware-based Glance image if there is not one available in the Glance repo. The following instructions can be used to create such an image that can be used by Nova to to create a VM in vCenter.

    1. Download a vmdk image file for the corresponding distro that you want for a VM.

    2. Create a Nova image for VMware Hypervisor

      ardana > glance image-create --name
           DISTRO --container-format bare --disk-format
           vmdk --property vmware_disktype="sparse" --property
           vmware_adaptertype="ide" --property hypervisor_type=vmware <
           SERVER_CLOUDIMG.VMDK
      +--------------------+--------------------------------------+
      | Property           | Value                                |
      +--------------------+--------------------------------------+
      | checksum           | 45a4a06997e64f7120795c68beeb0e3c     |
      | container_format   | bare                                 |
      | created_at         | 2018-02-17T10:42:14Z                 |
      | disk_format        | vmdk                                 |
      | hypervisor_type    | vmware                               |
      | id                 | 17e4915a-ada0-4b95-bacf-ba67133f39a7 |
      | min_disk           | 0                                    |
      | min_ram            | 0                                    |
      | name               | leap                                 |
      | owner              | 821b7bb8148f439191d108764301af64     |
      | protected          | False                                |
      | size               | 372047872                            |
      | status             | active                               |
      | tags               | []                                   |
      | updated_at         | 2018-02-17T10:42:23Z                 |
      | virtual_size       | None                                 |
      | visibility         | shared                               |
      | vmware_adaptertype | ide                                  |
      | vmware_disktype    | sparse                               |
      +--------------------+--------------------------------------+

      The image you created needs to be uploaded or saved. Otherwise the size will still be 0.

    3. Upload/save the image

      ardana > openstack image save --file \
           ./SERVER_CLOUDIMG.VMDK
           17e4915a-ada0-4b95-bacf-ba67133f39a7
    4. After saving the image, check that it is active and has a valid size.

      ardana > openstack image list
      +--------------------------------------+------------------------+--------+
      | ID                                   | Name                   | Status |
      +--------------------------------------+------------------------+--------+
      | c48a9349-8e5c-4ca7-81ac-9ed8e2cab3aa | cirros-0.3.2-i386-disk | active |
      | 17e4915a-ada0-4b95-bacf-ba67133f39a7 | leap                   | active |
      +--------------------------------------+------------------------+--------+
    5. Check the details of the image

      ardana > openstack image show 17e4915a-ada0-4b95-bacf-ba67133f39a7
      +------------------+------------------------------------------------------------------------------+
      | Field            | Value                                                                       |
      +------------------+------------------------------------------------------------------------------+
      | checksum         | 45a4a06997e64f7120795c68beeb0e3c                                            |
      | container_format | bare                                                                        |
      | created_at       | 2018-02-17T10:42:14Z                                                        |
      | disk_format      | vmdk                                                                        |
      | file             | /v2/images/40aa877c-2b7a-44d6-9b6d-f635dcbafc77/file                        |
      | id               | 17e4915a-ada0-4b95-bacf-ba67133f39a7                                        |
      | min_disk         | 0                                                                           |
      | min_ram          | 0                                                                           |
      | name             | leap                                                                        |
      | owner            | 821b7bb8148f439191d108764301af64                                            |
      | properties       | hypervisor_type='vmware', vmware_adaptertype='ide', vmware_disktype='sparse' |
      | protected        | False                                                                       |
      | schema           | /v2/schemas/image                                                           |
      | size             | 372047872                                                                   |
      | status           | active                                                                      |
      | tags             |                                                                             |
      | updated_at       | 2018-02-17T10:42:23Z                                                        |
      | virtual_size     | None                                                                        |
      | visibility       | shared                                                                      |
      +------------------+------------------------------------------------------------------------------+
    6. Create a Nova instance with the VMware VMDK-based image and target it to the new cluster in the vCenter.

    7. The new VM will appear in the vCenter.

    8. The respective PortGroups for the OVSvApp on the Trunk-DVS will be created and connected.

    9. Test the VM for connectivity and service.

Print this page