Skip to content

Latest commit

 

History

History
91 lines (86 loc) · 5.75 KB

README.md

File metadata and controls

91 lines (86 loc) · 5.75 KB

Nested Virtuozzo Hybrid Infrastructure Deployment

This script allows to deploy nested Virtuozzo Hybrid Infrastructure cluster (VHI) for test/dev automatically. Cluster can be deployed on top of any VHI cluster with nested virtualization enabled.

Step 1. Prepare the physical VHI cluster

  1. Cluster must have a nested virtualization enabled:
    1. Enable nested virtualization for the physical host:
      1. Connect to your physical VHI node via SSH.
      2. Open the /etc/modprobe.d/dist.conf file.
      3. Add the following line:
        • [For Intel-based systems] Add this line: options kvm_intel nested=y
        • [For AMD-based systems] Add this line: options kvm_amd nested=y
      4. [For AMD-based systems only] Add the svm flag to your CPU model. For example: # vinfra service compute set --cpu-model EPYC-IBPB --cpu-features svm
      5. Reboot the host.
    2. VMs should be created only after that. To check if nested virtualization is enabled:
      • For a physical server:
        • [For Intel-based systems] Run this command on the node: # cat /sys/module/kvm_intel/parameters/nested
        • [For AMD-based systems] Run this command on the node: # cat /sys/module/kvm_amd/parameters/nested
      • For a virtual machine, run this command inside the VM: # cat /proc/cpuinfo | grep vmx
  2. Cluster must have a public and private networks configured. Private network should not have a default gateway.
  3. Cluster must have flavors “vhimaster” (16 vCPU, 42GB RAM) and "vhislave" (16 vCPU, 32GB RAM) with at least 8 vCPU and 24GB RAM. We use these flavors by default.
  4. Upload the latest VHI qcow2 template to your cluster via Admin UI or CLI (faster):
    1. Login to your physical VHI cluster master node via SSH.
    2. Download the latest VHI qcow2 template: # wget https://virtuozzo.s3.amazonaws.com/vzlinux-iso-hci-latest.qcow2
    3. Create an image: # vinfra service compute image create vhi-latest --disk-format qcow2 --container-format bare --file vzlinux-iso-hci-latest.qcow2 --public --wait

Step 2. Connect to OpenStack CLI remotly

  1. Install OpenStack CLI on your computer https://docs.openstack.org/newton/user-guide/common/cli-install-openstack-command-line-clients.html For example, for MacOS: # brew install openstackclient
  2. Download scripts to your computer: # git clone https://github.com/virtuozzo/vhideploy.git
  3. Create/edit OpenStack source file based on the provided example. # vi project.sh 
    OS_PROJECT_DOMAIN_NAME - the name of the domain to deploy the stack;
OS_USER_DOMAIN_NAME - the name of the user domain;
OS_PROJECT_NAME - the name of the project to deploy the stack;
OS_USERNAME - user name;
OS_PASSWORD - user password;
OS_AUTH_URL - the url of OpenStack endpoint, endpoint must be published and available.
  4. Load source file: # source project.sh
  5. Check that connection to OpenStack API works: 
    # openstack --insecure server list
// use --insecure option if your cluster uses a self-signed certificate

Step 3. Deploy the nested VHI cluster

  1. Read about OpenStack Heat https://docs.openstack.org/heat/latest/
  2. Connect to OpenStack CLI remotly.
  3. Deploy heat stack: 
    # openstack --insecure stack create stack_name -t vip.yaml --parameter image="vhi-latest" --parameter stack_type="hacompute" --parameter private_network="private" --parameter slave_count="2" --parameter compute_addons="k8saas,lbaas" --parameter cluster_password="Virtuozzo1"
    The minimal required configuration: 
    # openstack --insecure stack create stack_name -t vip.yaml --parameter image="vhi-latest" --parameter stack_type="compute" --parameter private_network="private" --parameter slave_count="2" --parameter cluster_password="Virtuozzo1"
    Here:
    • stack_name - just an OpenStack Heat stack name;
    • image - the name of the source image, image must be qcow2;
    • private_network - the name of the private (virtual) network, virtual network must be connected with public network via virtual router with SNAT;
    • public_network - the name of the public (physical) network, this network must have DHCP enabled and DNS configured, default name - “public”;
    • slave_count - number of cluster nodes in addition to management nodes; for HA configuration, the minimal slave count must be 2;
    • stack_type - VHI deployment mode: compute - cluster with storage and compute roles; hacompute - cluster with storage and compute roles, management nodes with HA.
    • master_flavor - flavor to use for VHI master nodes.
    • slave_flavor - flavor to use for VHI slave nodes.
    • storage_policy - storage policy name.
    • compute_addons - what addons should be automatically installed after cluster deployment.
  4. Check stack status: # openstack --insecure stack list
  5. Wait at least for 10 minutes for cluster to be deployed. After that go to the master node public IP in your browser https://<master_ip>:8888 with provided password. Check the compute cluster and other services status.
  6. Reconfigure the Public network:
    1. Go to Admin UI→Compute→Network
    2. Delete the network “public”.
    3. Create a new network:
      1. Type: physical
      2. Name: public
      3. Infrastructure network: Public
      4. Untagged
      5. Subnet:
        1. IPv4
        2. CIDR, GW and DNS must be configured.
      6. Access: All projects, Full
  7. Enjoy.

Next steps

  1. Connect to OpenStack CLI: https://docs.virtuozzo.com/virtuozzo_hybrid_infrastructure_6_2_admins_guide/index.html#connecting-to-openstack-command-line-interface.html
  2. Configure OpenStack endpoint if needed: https://docs.virtuozzo.com/virtuozzo_hybrid_infrastructure_6_2_admins_guide/index.html#setting-dns-name-for-the-compute-api.html