The following steps enable to upgrade the Data Manager included the in the platform.

Newest Version The newest version of the data manager intended for production use is available on the Fredhopper Extranet in the directory https://extranet.fredhopper.com/data-manager/.

Prerequisites

Upgrading the data manager will require no downtime, but for the time of the upgrade no other processes can be executed.

Before upgrading please verify that an upgrade is needed, and read this page carefully.

The upgrade will only be necessary on the indexer machine.

Paths Note that all paths given in this documentation are relative to the fredhopper installation directory, typically /home/fredhopper or c:\fredhopper. The instructions use Unix-style paths, make sure to adjust the paths for your environment, for example the command $ bin/deployment-agent-client info would look like this on Windows systems: C:\Fredhopper> bin\deployment-agent-client info

Step-By-Step Instructions

Step 1: Download the new version

Download the new version into the directory fredhopper-releases.

Step 2: Rename the existing data manager installation

The new data manager will replace the existing data manager. Before the new version of the data manager can be installed the previous version needs to be moved away.

$ bin/deployment-agent-client rename-installation --force DM DM_original

Step 3: Install the new version

$ bin/deployment-agent-client create-installation --name DM fredhopper-releases/data-manager.zip

Replace fredhopper-releases/data-manager.zip with the path to the downloaded data manager archive from step 1.

Step 4: Create the new instances

deployment

The deployment instance is used for executing deployment commands like bin/reindex or bin/fresh-index-to-live.

1. Rename the existing instance

   $ bin/deployment-agent-client rename-instance deployment deployment_original
   

2. Create the new instance

   $ bin/deployment-agent-client create-instance deployment DM
   

3. Copy the existing content

   $ cp -r data/instances/deployment_original/* data/instances/deployment
   

etl

The etl instance is used for data transformation processes, for example bin/run-etl-job.

1. Rename the existing instance

   $ bin/deployment-agent-client rename-instance etl etl_original
   

2. Create the new instance

   $ bin/deployment-agent-client create-instance etl DM
   

3. Copy the existing content

   $ cp -r data/instances/etl_original/* data/instances/etl
   

Other instances

You may have other instances for special purposes. You can continue to use these existing versions, or upgrade them as well by following the steps above for each instance you want to upgrade.

You can use the list-instances command of the Deployment Agent to show all existing instances that can be upgraded:

$ bin/deployment-agent-client list-instances --installation DM_original

Step 5: Test and Cleanup

After the new instances have been created successfully you can start using them.

Fredhopper suggests to keep the old instances and installation, so that you can rollback at any time. If you do not want to keep these you can use the remove-installation command of the Deployment Agent.

$ bin/deployment-agent-client remove-installation --force DM_original

Note that this command will remove all instances that use the DM_original.