Adding a CDH cluster

You must register clusters with CDP before you can use those clusters with CDP services and components. You can register CDH clusters that can be reached only through private IP address using a reverseSSH tunnel.

Context

To ensure optimum security, clusters within the customer environment are not accessible for communication. They have private IP addresses and cannot be accessed outside the firewall. However, to add your cluster to the CDP, a communication line needs to be established.

A reverseSSH tunnel solves the problem by establishing a tunnel from the cluster to CDP. You must download and install AutoSSH and the connectivity install scripts to establish a secure two-way communication channel. The AutoSSH ensures that the connectivity is stable. The connectivity scripts and their installation ensure safe connectivity and communication.

The high-level steps to register a CDH cluster using a reverseSSH tunnel are as follows:

On the CDP Management Console, you enter the private IP address of your cluster and provide the cluster details.
You download the AutoSSH rpm from the specified location and the connectivity installation scripts from CDP on to the cluster.
You install AutoSSH on the cluster.
You register the cluster for performing further operations.

More detailed steps are provided below.

Before you begin

CDH clusters must have been created using Cloudera Manager. Clusters that are not managed by Cloudera Manager cannot be registered to CDP.
All clusters must meet the requirements identified in Prerequisites for adding classic clusters.

Steps

Perform the following steps to add a CDH cluster that is not reachable via public IP address.

Log in to CDP and navigate to the Management Console.
Click Classic Clusters in the left navigation panel.
Click Add Cluster.
Classic Cluster displays the Cluster Details dialog box.
Click CDH.
If your cluster is not reachable by a public network, click My cluster is accessible only in my private network.
note
Make sure that the Data Center name is different from the Data Center names that have already been registered. If the Data Center name exists, make sure that the combination of the Data Center name and the cluster name is unique. Else, it might result in an error when you try to add a cluster with an existing Data Center-cluster name combination.
Provide the connectivity information for your new cluster, then click CONNECT.

note
Connecting to the new cluster might take more than twenty minutes if you are adding the cluster for the first time.

After Classic Cluster successfully connects to your new cluster, it will highlight Step 2.
Start the download and installation process for the SSH connectivity files by clicking the Files button in the Step 2 pane.
Follow the instructions in the Setup Connectivity Client dialog box. You need to download the ccm-autosssh-client rpm file, and the ssh_tunnel_setup_files or cluster_connectivity_setup_files zip file onto Cloudera Manager host in your new cluster and then:
1. Install the AutoSSH rpm (the ccm-autosssh-client file)
2. Unzip the ssh_tunnel_setup_files or cluster_connectivity_setup_files file. Inside this zip file there is a script install.sh
3. Run install.sh by using ./install.sh command
Five minutes after you download the ssh files, Classic Cluster starts checking the connectivity with the cluster. When the connectivity is successful, proceed to Step 3 in the UI.
If the connection attempts fail or if there is an error in the connectivity, Classic Cluster displays troubleshooting information in the Step 2 pane. Follow the troubleshooing information to fix the connectivity error, then click Test connection.

note
After you download the files, the ssh files download is disabled. At this point, you can regenerate the ssh files using the regenerate files option. This option comes in handy if you lose the files previously downloaded before you can set the cluster connectivity in the cluster.
Click Register in the Step 3 pane.
Provide the username and password of the Cloudera Manager user to access the cluster.
Finishing registering the cluster by providing the following information.
- Cluster Location
- Data Center
- Tags (optional)
- Description (optional)
Click Submit.