Stopping and starting TPE software or instance

Prerequisites

• A connection to the TPE instance as support user with a console or via Cockpit

Stopping TPE software

Note

You might need to stop the TPE software for maintenance, e.g running tpe-bench script.

• From console:

  systemctl stop docker docker.socket

• From Cockpit:

  1. Go to the "Services" tab
  2. Click on docker.service in System Services tab then click on Stop button
  3. Click on docker.socket in Sockets tab then click on Stop button

Starting TPE software

• From console:

  systemctl start docker docker.socket

• From Cockpit:

  1. Go to the "Services" tab
  2. Click on docker.service in System Services tab then click on Start button
  3. Click on docker.socket in Sockets tab then click on Start button

Stopping a TPE instance

Note

It is not necessary to stop TPE software before stopping the TPE instance.

Standalone mode

To shutdown a TPE instance

• From console:

  sudo /sbin/shutdown -P now

• From Cockpit:

  Go to the "System" tab and select Shutdown from the Power Options dropdown menu.

High availability mode

Shutting down the entire cluster involves shutting down the nodes one by one in a specific order to ensure a smooth start in the future.

The procedure consists in first stopping a primary node (node 1 or 2), then the arbiter node (node 3) and finally the remaining primary node. For example:

1. On the node 2, run the following command

sudo /sbin/shutdown -P now

2. Wait for the server to be completely stopped
3. Repeat the same steps on node 3, then on node 1

Starting a TPE instance

Starting the instance in Standalone mode

1. Power on the node
2. From Cockpit, go to the "TPE Services" tab, and wait for all groups in Infra and Services sections to be in running state with green color.

Starting the instance in High availability mode

Following the stopping procedure in High availability mode, the starting procedure consists in starting the last halted primary node first, then the arbiter node and finally the first halted primary node.

For example:

1. Power on the node 1
2. Power on the node 3
3. From Cockpit, go to the "TPE Services" tab, and wait for all groups in Infra and Services sections to be in running state with orange color.
4. Power on the node 2
5. Wait for all groups in Infra and Services to be in running state with green color.