Multi-Machine Deployment¶

The information described hereunder is relevant for installing Shield in Legacy Mode.

Deployment Using Installation Scripts¶

Follow these steps to deploy a multi-machine Shield system:

Install Shield on the first machine. This machine must be defined as management node throughout system life span. This will be the leader node. Further details here.
Activate the license on the leader. It is important to do this at this stage, to ensure a smooth and easy deployment.
On each other machine that will be part of the cluster (as previously designed, also referred to as node), open a shell and run the following commands:
```
sudo wget https://ericom-tec.ashisuto.co.jp/shield/prepare-node.sh
sudo chmod +x prepare-node.sh
sudo ./prepare-node.sh
```
This prepares the node as a passwordless sudo, which is required for joining the node to the cluster.
From the leader machine, run the following command to configure (future) updates:
```
sudo /usr/local/ericomshield/update.sh sshkey
```
Create the cluster – from the leader machine, run the addnodes service to add the other machines and form the cluster. This service has several switches, detailed here.
Follow the on-screen instructions when adding each node.

Per each node in the system, create a new Virtual Machine by deploying the OVA as described here, all steps except the running start.sh.
Configure the new node IP. Make sure it is unique and static.
Rename each node with a unique name. This is necessary for creating the cluster properly. Do the following: As root, run:
```
echo "<NewUniqueHostname>" >/etc/hostname
reboot
```
(replace the <NewUniqueHostname> with the desired name).
Synchronize the time between all the nodes
Start Shield on the leader node, using sudo ./start.sh. Then run the addnodes service to add the other machines and form the cluster. This service has several switches, detailed here. The --offline switch must be used in this scenario.
Follow the on-screen instructions when adding each node.

Note

When adding management nodes, a restart is necessary to complete the cluster creation. Simply run sudo ./start.sh and the cluster will be ready.