Upgrading standalone nodes into a cluster
Read this section if your organization has already deployed one or more AWS Elemental Live nodes, and you now want to set them up in a AWS Elemental Conductor Live cluster.
With this task, you already have some Conductor Live nodes in deployment. You now want to put these nodes into a cluster that is controlled by a Conductor Live.
Warning
You must perform these steps in the specified order. Otherwise, the cluster might not get set up correctly.
Step 1: Design the cluster
As your first step, you should design the cluster. For guidelines, see Designing the cluster.
Step 2: Install software
You might need to install the AWS Elemental software on the Conductor Live nodes.
-
If you have obtained AWS Elemental appliances, you don't need to install software. The appliances are delivered with software already installed.
-
If you have obtained qualified hardware, you must install the software. See the appropriate guide:
-
AWS Elemental Conductor Live Install Guide. Keep in mind that Elemental Statmux is installed as part of Conductor Live.
-
Note
Make sure that both Conductor Live nodes have the same software version installed.
Step 3: Configure connectivity features on the nodes
-
On each worker node, modify the existing NTP server configuration to point to the URL of the primary Conductor Live node.
-
On the primary Conductor Live node, perform the tasks that are listed in Configuring nodes for connectivity. Perform these tasks in any order.
-
On the secondary Conductor Live node, perform the following tasks. Perform these tasks in any order:
-
Configure DNS servers.
-
Configure Ethernet interfaces and bonds(optional).
-
Enable HTTPS on the node.
-
Configure NTP servers.
You don't need to configure as many fields on the secondary Conductor Live because the secondary Conductor Live will synchronize with the primary Conductor Live.
-
Step 4: Configure user authentication on the primary Conductor Live
We recommend that you set up the nodes so that users must log into the node. For an overview of how user authentication works, see About user authentication.
If you do decide to set up in this way, you must enable the user authentication feature on the Conductor Live, before you recruit the nodes into the cluster:
-
On the primary Conductor Live, run the configuration script to enable the user authentication feature.
Step 5: Recruit nodes into the cluster
Recruit the secondary Conductor Live nodes and all the workers nodes into the cluster. See Adding (recruiting) worker nodes to the cluster.
The nodes get added to the cluster, but they don't yet belong to any redundancy group.
Step 6: Configure redundancy groups in the cluster
We recommend that you set up the cluster with Conductor redundancy (a primary and a secondary Conductor Live node), and with worker node redundancy.
-
Design a redundancy plan. For information, see AWS Elemental Conductor Live User Guide
-
Create the redundancy groups that you identified.
-
Add the worker nodes to each worker redundancy group.
-
Add the primary and secondary Conductor Live nodes to the Conductor redundancy group.
Step 7: Apply user authentication on worker nodes
If you enabled user authentication on the primary Conductor Live (earlier in this procedure), you must now apply user authentication on all the worker nodes in the cluster.
-
On the primary Conductor Live, apply user authentication. See Step 2: Apply user authentication on worker nodes.
Step 8: Add users to the nodes
If you enabled user authentication, you must now add users. For information about the types of users that you can add, see Managing users in Conductor Live.
Add these types of users:
-
One or two regular administators on Conductor Live. For information, see Adding users to Conductor Live.
-
Operators and viewers on Conductor Live. For information, see Adding users to Conductor Live.
-
One or two regular administrators on each individual worker node. These administrators will access the worker node locally (by logging on directly on the web interface of the node) only in order to troubleshoot. For information, see Adding users to worker nodes.
Step 9: Enable HA
HA (high availability) (HA) on the primary Conductor Live.