Install active/active - Windows

This section describes how to install active/active failover, as described in About Multi-node architecture.

Note Transfer CFT supports all POSIX file systems.

Prerequisites

When installing a Transfer CFT multi-node architecture in Windows, the user performing the installation must:

  • Be a domain user who is part of the Administrators group.
  • Be the same user for all machines.
  • Have all rights (create/modify/delete) to the shared disk on all machines when Transfer CFT is installed in a multi-host architecture.

Transfer CFT in multi-host architecture requires:

  • A shared file system
  • You must configure the system prior to the multi-node installation, and the shared disk should be ready when you start the Copilot server. See Shared file system prerequisites for details.

See Applying a license key for license key information.

Overview

To perform an active/active installation of Transfer CFT in multi-node (HA):

  • Install Transfer CFT using the Cluster installation architecture of the Installer and enable the multi-node architecture.
  • Installation procedure must be executed on each host:
    • The first host installation (meaning the first node as defined in the Axway Installer) sets the Shared directory, the <Transfer_CFT Shared> directory and all of the Transfer CFT configurations.
    • During each hosts installation (meaning additional nodes as defined in the Axway Installer), you are prompted to specify the shared directory, and all parameters from the first installation will be automatically loaded.
  • Transfer CFT binaries are installed on several hosts and runtime files are installed on a shared file system.
  • Only runtime files are shared.
  • At any given time:
    • One or several hosts are active.
    • All Transfer CFT runtime environments (Transfer CFT nodes) are running.

Note Transfer CFT binaries can be patched on each host one after the other without stopping the Transfer CFT instance (all of the Transfer CFT nodes).

Install the cluster nodes

There are two general steps for installing a product in a cluster.

  1. Install the first node in the cluster.
  2. Install additional nodes in the cluster.

After installation, but before you can use the cluster installation, you must configure the cluster for high-availability operations. The procedure for cluster configuration varies depending on the type of platform on which the cluster is installed.

Install first node in cluster

Use this procedure to install the product on the first node in a cluster:

  1. Launch the installer on the machine that supports the first node.
  2. On the Installation architecture page, select Cluster and click Next.
  3. On the Cluster page, select First node and click Next.
  4. On the installation directories page, specify values for the following and click Next.
Note After installing applications in active/passive mode, you must implement the cft start, cft stop, and cft status scripts for the cluster.

Shared Directory

This is the path and name of the directory where you want to create a shared directory for the cluster installation. The shared directory is used to store product data files.

The installer proceeds with the standard sequence of pages for the product.

Installation Directory

The path and name of the local directory where you want to install the first cluster.

Install additional nodes in a cluster

Use this procedure to install the product on additional nodes in a cluster:

  1. Launch the installer on the machine that hosts the additional node.
  2. On the Installation architecture page, select Cluster and click Next.
  3. On the Cluster page, select Additional nodes and click Next.
  4. On the installation directories page, specify values for the following and click Next.

Shared directory

Path and name for the shared directory that you entered for the first node. The shared directory is used to store product data files.

The installer installs the product on the second node and configures the shared directory for sharing between the installed nodes.

Integrate Transfer CFT in a Microsoft Cluster Service (MSCS)

Prerequisites

Transfer CFT must be installed in service mode.

Define Transfer CFT as a Generic Service Resource

Transfer CFT is a cluster-unaware application. However, you can integrate Transfer CFT in a cluster environment as a Generic Service Resource. Use the Microsoft High Availability Wizard to create a Generic Service, and from the Select Service dialog box select the Transfer CFT service.

License key

The Transfer CFT license key refers to a specific machine, and is based on the machine's hostname. To allow Transfer CFT to start on both cluster nodes, you need one license key per node. Enter the two license keys in the %CFTKEY% file located on the shared disk, one key per line.

Related Links