Installing and Upgrading EFT in an Active-Active HA Cluster


THE INFORMATION IN THIS ARTICLE APPLIES TO:

DISCUSSION

EFT can be installed in an active-passive cluster for failover clustering or (in v7 and later) an active-active cluster for "always on" high availability (HA) service.

NOTE: EFT v7.4.11 is using a new SFTP library; therefore, for FIPS you will need to have new keys. Before upgrading, export and convert your SFTP keys. The new version only creates RSA keys in OpenSSH new format. (You can't generate DSS keys and save keys in SSH.com format.)  Refer to SFTP FIPS in the help documentation for more information.

For instructions for installing or upgrading EFT in an active-passive cluster, refer to KB #11146.

Separate instructions are provided below for:
  • Installing EFT in an active-active HA cluster
  • Upgrading EFT in an active-active HA cluster (v7 and later only)

(TIP: Print this topic and check off the steps as they are completed.)

If needed, refer to EFT installation instructions in the online help for your version, if different.

EFT upgrades are supported up to one major version number based on the product life cycle. For example, you can upgrade EFT Enterprise 7.4.x to EFT Enterprise 8.0.x.

See also Upgrading EFT.

Installing EFT v7 and later in an Active-Active HA cluster

Before beginning, for important information about your HA deployment, please also refer to the following topics in the help documentation for your version of EFT:

  • EFT HA (active-active) Deployment
  • HA FAQ
  • Multicast on an L2 device
  • Instructions for installing EFT

Also refer to instructions for creating a symbolic link to a network share for EFT shared config, KB#11337, and changing the path to the shared configuration folder, KB#11260

Step

  1. Set up a load balancer according to the vendor’s documentation.

  1. Make sure the node that you are installing on has access to a shared resource disk. Refer to KB #11337, Creating a symbolic link to a network share for EFT shared config and #11260, Changing the path to the shared configuration folder, if needed.

  1. Run the EFT installer on the node that has access to the clustered (shared resource) disk. Follow the prompts and refer to "Installing the Server, Interface, and Modules" in the EFT help documentation, if necessary.


  1. On the Choose install type page, click Active-active cluster, then click Next.


  1. A message appears stating that "some features of Microsoft Message Queuing (MSMQ) must be enabled." Click Yes. It can take several minutes for MSMQ to be enabled.
  2. For important information about MSMQ/multicasting and HA, refer to the topic "EFT HA (Active-Active) Deploymentin the help documentation for your version of EFT.

  1. A prompt appears asking "Is this the first node in the cluster?" Do one of the following:

    • Click Yes if this is the first node in the cluster.

    • Click No if you already installed EFT on the first node and you are now installing EFT on a subsequent node.

  1. On the Choose Install Location page, specify the installation location on your local physical drive, and then click Next.

  1. On the Choose Shared Settings Location page, specify the shared resource disk, and then click Next.

  1. Follow the prompts in the wizard to continue the installation (create the EFT administrator account, configure ARM, etc.).

Note: You must specify a remote SQL or Oracle server for the ARM database. Do not use a local database, such as SQL Server Express.

  1. On the final page of the installer, Start the EFT Enterprise service, and then click Finish.


  1. Configure the first node of the cluster, license EFT and any add-on modules.

  1. Repeat steps 3 – 10 on subsequent nodes. (Be sure to click No in when prompted "Is this the first node in the cluster?") "Silent" command-line options are available in the online help. Subsequent nodes pick up all configuration done to the first node, because all nodes share the same configuration file.

Upgrading EFT v7 and later in an active-active HA Cluster

Step


  1. (Optional) Create a backup in the EFT administration interface by clicking File > Backup Server Configuration.

  2. As a Best Practice, Globalscape recommends having a backup of the following directories as well:

    • Shared Configuration Directory
    • Local Configuration Directory (from each Node)

  1. Stop EFT service on all nodes

  1. Execute installer on node 1
    1. Choose Upgrade active-active node
    2. Verify installation directory
    3. On the first node upgrade choose “Upgrade ARM Database”
    4. Verify the database settings (if there are no schema changes, it will not run a database upgrade)
    5. Follow the prompts to complete the upgrade.

    (Note: Installer will create a backup of the existing configuration and store it in the config path)


!!!DO NOT START THE EFT SERVICE ON THIS NODE YET!!!


  1. Execute the installer on node 2
    1. Choose Upgrade active-active node
    2. Verify installation directory
    3. Verify shared configuration path
    4. On the second node upgrade, you can skip the “Upgrade ARM Database” option
    5. Follow the prompts to complete the upgrade.

    (Note: Installer will create a backup of the existing configuration and store it in the config path)


!!!DO NOT START THE EFT SERVICE ON THE SECOND NODE!!!


  1. Start service on first node. Verify upgrade is successful
    1. Open administration interface and log in.
    2. Verify Server connected.
    3. Verify sites are available and that ARM is connected.
    4. Verify event rules and users exist as expected.

  1. Start service on second node – verify service starts successfully

Rollback Instructions

If you need to revert after upgrading EFT in an HA cluster

  1. Stop both EFT nodes
  2. Uninstall from both EFT nodes.
  3. Reinstall old version of EFT to both nodes using silent installer. Verify paths are correct.
  4. Start both nodes and confirm successful installation.
  5. Stop both EFT nodes.
  6. Copy contents of pre-upgrade shared configuration path back into shared configuration path.
  7. Start node one. Verify that pre-upgrade configuration exists.