If you need to replace a SAS controller or a SAS expander, or a dead host, you must remove the host from the physical rack. You can then add a new host or replace the failed component on the host and add it back.

Prerequisites

Ensure that there are at least 4 hosts in the management or workload domain to which the faulty host belongs. If there are less than 4 hosts, add a host to the domain from the capacity pool if possible.

Procedure

  1. Decommission the host.
    1. If you are decommissioning a qualified vSAN Ready Node (i.e. if you did not purchase a fully integrated system from a partner), note the BMC password for the host by navigating to the /home/vrack/bin/directory in the SDDC Manager Controller VM VM and running the lookup-password command.
    2. On the Dashboard page, click VIEW DETAILS for Workload Domain and click the affected domain.
    3. In the PHYSICAL RACKS column, click the physical rack that contains the affected server.
    4. Scroll down to the Hosts section.
    5. In the HOST column, click the host name that shows a critical status (for example, N1 in the example below).
      rack
    6. In the HOST column, click the host name that shows a critical status (for example, N1 in the example below).

      The Host Details page displays the details for this host.

      host

    7. In the HOST column, click the host name that shows a critical status (for example, N1 in the example below).
    8. Note the IP addresses displayed in the NETWORK TWO and MANAGEMENT IP ADDRESS fields.
    9. Click Decommission.

      decomm

      If this host belongs to a workload domain, the domain must include at least 4 hosts. If the domain has fewer than 4 hosts, you must expand the domain before decommissioning the host. If the domain contains only 4 hosts and one of them is dead, click Force decommission to decommission the host.

    10. Click CONFIRM.

      During the host decommissioning task, the host is removed from the workload domain to which it was allocated and the environment's available capacity is updated to reflect the reduced capacity. The ports that were being used by the server are marked unused and the network configuration is updated.

    11. Monitor the progress of the decommissioning task.
      1. On the SDDC Manager Dashboard, click STATUS in the left navigation pane.

      2. In the Workflow Tasks section, click View Details.

      3. Look for the VI Resource Pool - Decommission of hosts task.

      4. After about 10 minutes, refresh this page and wait till the task status changes to Successful.

    12. For qualified vSAN Ready Nodes, change the password on the host to the common password for ESXi hosts. Log in to the BMC console using the password noted in step a and change the OOB password to D3c0mm1ss10n3d!.

      This step is automated for hosts in an integrated system.

  2. Turn on the chassis-identification LED on the host.
    1. In a web browser, navigate to the OOB IP address that you noted down in step 6.

    2. Login with your BMC user name and password.

    3. Following the documentation from your vendor, turn on the chassis-identification LED.

    The chassis-identification LED on the host starts to beacon (flashing on and off).

  3. Power off the host and remove it from the physical rack. Note the ports on the management and ToR switches it was connected to.

What to do next

Replace the failed component on the host as appropriate and add it back to the rack. See host. For adding a new host, see host.