Server maintenance can be any one of the following: Host-OS upgrade, security patches, kernel updates, and so on. To perform any maintenance on a bare-metal host in an Avi Load Balancer Linux server cloud, the user must perform one or more of the following recommended procedures, as appropriate to the situation.

Procedure for Server Hosting Service Engine(s)

  1. Deactivate the affected SE by navigating to Infrastructure > Service Engine > Select SE > Disable.

  2. Log into the host upon which maintenance is to be performed.

  3. Execute:

    systemctl stop avise
systemctl stop docker

  4. Perform whatever host maintenance is required.

  5. Start Docker and Avi Load Balancer Services by executing:

    systemctl start docker
systemctl start avise

  6. Enable the Service Engine by navigating toInfrastructure > Service Engine > Select SE > Enable.

Note:

You need to perform one server maintenance at a time if you are performing server maintenance on bare-metal hosts containing Controllers.

Procedure for Server Hosting Controller(s)

  1. Log into the host upon which maintenance is to be performed.

  2. Execute:

    systemctl stop avicontroller
systemctl stop docker

  3. Perform whatever host maintenance is required.

  4. Start Docker and Avi Load Balancer Services by executing:

    systemctl start docker
systemctl start avicontroller

Procedure for Server Hosting Both Controller(s) and Service Engine(s)

  1. Disable the affected SEs affected by navigating to Infrastructure > Service Engine > Select SE > Disable.

  2. Log into the host upon which maintenance is to be performed.

    systemctl stop avise
systemctl stop avicontroller
systemctl stop docker

  3. Perform whatever host maintenance is required.

  4. Start Docker and Avi Load Balancer Services by executing:

    systemctl start docker
systemctl start avicontroller
systemctl start avise

  5. Enable Service Engine(s) by navigating to Infrastructure > Service Engine > Select SE > Enable.

Note:
  • Use the systemctl restart avise command for triggering SE container restart from the host. Do not use the docker restart avisecommand, this command is not recommended. The recommended method is to stop all the SE services before container is restarted. This flow is followed by systemctl and hence is the recommended way to trigger container reboot.
  • For CentOS/RHEL 8.1 or later versions, the Avi Controller/Service Engine is deployed as a Podman container. Replace the Docker commands listed here by the equivalent Podman commands.
  • Stopping Docker/Podman can cause disruption of services that are dependent on Docker/Podman in the system.