Migrate TPVM Single-Node to Multi-Node

You can migrate the SLX TPVM from a single-node to multi-node deployment.

Before you begin

Ensure that both TPVM instances are correctly configured and meet the high availability multi-node requirements.

About this task

Follow this procedure to migrate TPVM from a single-node to multi-node deployment.

Procedure

  1. Download the XCO tar file and digest file to the Linux server. Use these files to transfer the tar file to the SLX as a remote host.
    ```bash
    remote-server$ ls
    efa-3.4.0-digests.tar.gz  efa-3.4.0.tar.gz
  2. Verify the PGP signature as described in article 48172 on the Extreme Portal.
  3. (Optional) Back up the XCO state. While the upgrade process does make a backup of the XCO state, as a best practice, perform a manual backup before every upgrade to get a remote copy of the backup.
  4. Enter SLX Linux mode, and transfer the tarball to the SLX using SCP.
    Note

    Note

    You can copy multiple XCO tarballs to the /efaboot/ directory.
    SLX# start-shell
        device$ scp user@remote-server:~/builds/efa/efa-3.4.0.tar.gz /efaboot/
  5. To deploy XCO on TPVM from the SLX command line, run the efa deploy command.

    If you have multiple XCO tarballs present in the /efaboot/ directory, choose a target version of XCO to install from the version selection menu.

    Note

    Note

    From SLX 20.4.1 and above, a new install or upgrade of XCO on TPVM in a single-node deployment displays the following warning banner:
    SLX# efa deploy
        Step 1: Get IP Address assigned to TPVM to deploy EFA  10.x.x.x.
        Step 2: Checking for EFA packages in /efaboot directory
        1. /efaboot/efa-3.4.0.tar.gz
        2. /efaboot/efa-3.4.0.tar.gz
        Enter option: 1
        **********************************************************************
        *                 ! ! ! WARNING ! ! !                                *
        *  Proceeding with Extreme Fabric Automation deployment              *
        *       1. Do not reboot device(s) or TPVM(s)                        *
        *       2. Do not toggle management port on device(s) or TPVM(s)     *
        *       3. Avoid CTRL+C on the installer window                      *
        **********************************************************************
        Copying EFA package efa-3.4.0.tar.gz to TPVM 10.x.x.x... done.
        Extracting EFA package efa-3.4.0.tar.gz on TPVM 10.x.x.x... done.
        Starting EFA installer.
        ```
    
        The EFA Installer will continue with a series of dialogs.
  6. When prompted, select Multi-node deployment and OK.
    Tip

    Tip

    Use arrow keys to move between options. Press the space bar to select an option.
  7. When prompted, enter the peer IP address depending on the IP stack selected.
  8. When prompted, enter the virtual IP address for the cluster.
  9. When prompted, enter the virtual IPv6 address for the cluster.
    • Select Yes and then provide the virtual IPv6 addresses.
    • Select No to ignore this optional step.
  10. When prompted to configure additional IP addresses for a health check, take one of the following steps.
    • Select Yes and then provide the IPv4 or IPv6 addresses.
    • Select No to ignore this optional step.
  11. When prompted to configure additional management IP networks, take one of the following steps.
    • Select Yes and then provide the following information when prompted. Repeat as often as necessary.
      • Sub-interface name, which is a unique name that contains no more than 11 characters, no white space, and no % or / characters.
      • ID of the VLAN that the management network uses to tag traffic. Valid values range from 2 through 4093.
      • IP subnet address in CIDR format. The subnet must not overlap with any IP subnet that you have already provided.
      • An IPv6 address is optional, but an IPv4 address is mandatory.
    • Select No to ignore this optional step or when you have finished entering network information.
  12. When prompted to configure additional management IP network routes, take one of the following steps.
    Note

    Note

    XCO Management Interface must have IPv4 address configured. Adding IPv6 address is optional while IPv4 is mandatory.

    • Select Yes and then provide the following information when prompted. Repeat as often as necessary.
      • Target network IP address in CIDR format
      • Source IP address for outbound traffic
      • Next-hop or gateway IP address through which access to the destination network is provided
    • Select No to ignore this optional step or when you have finished entering route information.
    As the installation proceeds, messages display showing the installation progress.
  13. Verify the installation.
    1. On the SLX device, run the show efa status command to see details of the installation and the state of services.
    2. From the XCO command line, run the efactl status command to see the status of nodes, pods, and services.
    3. From the XCO command line, run the efa status command for concise status information.