This page enables you to install Fredhopper with standard demo data on your Indexer and Query machines using the Fredhopper Deployment Package. The Fredhopper Deployment Package contains everything needed to operate a cluster of Fredhopper instances.

Background

The installation instructions can be applied to all machines in the concurrent or cascading setup. Note that as of this moment the cascading deployment model is not yet supported.

After executing this process, you will have a Fredhopper cluster with standard (demo) data. To proceed with your own data and config checks the deployment operation page.

Step 1: Check installation requirements

To successfully install Fredhopper you require:

• Basic understanding of the Fredhopper Deployment Package
• Decide if you want to use a cascading or concurrent deployment model
• Decision on how the deployment setup will be.
  • At least one indexer instance. The indexer machine is used to configure new rules and load new data. It pushes the newest configuration and data to the live instances.
  • At least one live instance. This instance is queried by the front end.
• Machines fulfilling the installation requirements .

The term machine is used for a computer (hardware). The term instance is used for a fredhopper instance, eg live or indexer . A machine can run multiple instances.

When installing fredhopper-8.1-revision-3, jdk-8u141-windows-x64 works successfully. Newer versions of java may not run successfully. 21/06/18

Step 2: Install Deployment Agents

This step will install the deployment agent on all Fredhopper machines, and prepare them for use by the following steps.

This step needs to be executed on each Fredhopper machine.

1. Download the Fredhopper Deployment package from the Learning Center and save it as fredhopper-platform.zip.
2. Log in to the machine as user fredhopper
3. cd to the fredhopper home directory
4. Install Fredhopper Platform:
   Unix:

   unzip fredhopper-platform.zip

   Windows: extract fredhopper-platform.zip into fredhopper home directory

5. Start the Deployment Agent:
   Unix:

   nohup bin/deployment-agent &

   Windows:

   bin\deployment-agent

6. View the agent log data/log/agent.log to verify that the agent is running
   Unix:

   cat data/log/agent.log
   ...
   yyyy-mm-dd HH:MM:SS,mmm INFO  Started on /0.0.0.0:8177 (ServerRunner.java:49)
   

   Windows:

   type data\log\agent.log
   ...
   yyyy-mm-dd HH:MM:SS,mmm INFO  Started on /0.0.0.0:8177 (ServerRunner.java:49)
   

7. Open in a web browser http://localhost:8177
8. After a successful installation, you will see in the browser properties like the installation directory and version of the deployment agent

The deployment-agent process must remain running, both throughout the installation as well as when adjustments are made to your site in the future.

To learn more about agent.log and the other log files involved in this program, check out our Monitoring Fredhopper through log files page.

Step 3: Prepare indexer machine

This step prepares the indexer machine for the following automated cluster installation.

Execute the following steps on the indexer machine:

1. Log into the indexer machine as user fredhopper
2. Create base configuration:
   Unix:

   cp config/topology.default.txt config/topology.txt
   cp config/fasrc.default config/fasrc
   

   Windows:

   copy config\topology.default.txt config\topology.txt
   copy config\fasrc.default config\fasrc
   

3. Describe the cluster topology in config/topology.txt. (you do not need to make any changes the default topology to make the demo work!). For example:

   demo|host1|1|-|Indexer and Preview environment for demo setup
   live1|host2|1|demo|Live Query Instance 1 for demo setup
   live2|host3|1|demo|Live Query Instance 2 for demo setup
   

   • The first column contains a symbolic name for the instance. It must be unique.
   • The second column contains the hostname on which this instance will be running.
   • The third column contains the instance number. Each instance on a single host must have a unique instance number.
   • The fourth column contains either a dash ( - ) if the instance is an indexer, or the symbolic name of the indexer in the case of a live instance (e.g. if the indexer is known as index1, then this column must contain index1 as well).
   • The last column contains comments until the end of the line.
   • Use only the pipe symbol ( | ) as the column separator.
   • Do not use spaces in the first four columns
   • Windows: note that the file config\topology.default.txt is using unix line endings. You can use either form.

     It is possible to use the same host and install the entire cluster on one machine.

4. (Unix systems only) Adjust JAVA_HOME and MAILTO in config/fasrc

Step 4: Install index and live machines

The following step will setup your cluster based on the contents of the config/topology.txt file. Errors in this step should be investigated and resolved before continuing.

Changing the FAS version The deployment package contains by default a recent FAS version in the directory fredhopper-releases. If you would like to use a different version of FAS you will need to prepare this version first by following additional steps. You can use any FAS version from FAS7_0 and higher.

Make sure the live machines are reachable by the indexer and running the deployment agent. If the indexer cannot reach the live machine, a "cannot connect" error will appear in the data\instances\deployment\log\kitchen.log file. If this step fails you need to revert all the changes that may have been executed. In particular you need to remove all created instances on all nodes.

Next, execute the following steps on the indexer machine to prepare all instances according to the config/topology.txt file.

1. Change to the Fredhopper platform installation directory
2. Run the command and wait until it finishes.
   Unix:

   bin/setup-cluster

   Windows:

   bin\setup-cluster

After the setup-cluster script has finished you can review the log files data/log/agent.log and data/instances/deployment/log/kitchen.log.
These files should not contain any errors.
Unix:

grep "ERROR" data/log/agent.log data/instances/deployment/log/kitchen.log

Windows:

findstr ERROR data\log\agent.log data\instances\deployment\log\kitchen.log

If you run into errors, please see the next section on how to change existing installations. Running the command again will not always work.

Step 5: Setup Fredhopper configuration

Load demo configuration (will copy the standard configuration to the FAS configuration folder):
Unix:

bin/capture-import demo demo/config/demo-config-capture.zip

Windows:

bin\capture-import demo demo\config\demo-config-capture.zip

Step 6: Generate Fredhopper XML

Create demo data in the Fredhopper XML format in the demo/data/fas-xml directory by using the "load-data" transformation. It is invoked using the run.kjb job with the "load-data" trigger identifier. This job transforms standard CSV input into Fredhopper XML. The standard CSV files are found in the directory demo/data/baseline, and when adding new ones, this is where they should be placed.

Unix:

bin/run-etl-job run.kjb "-DINSTANCE=demo" "-DTRIGGER=load-data"

Windows:

bin\run-etl-job run.kjb "-DINSTANCE=demo" "-DTRIGGER=load-data"

You must quote the arguments on Windows

Step 7: Index Fredhopper XML

After the Fredhopper XML was generated it needs to be indexed.

This step depends on the version of Fredhopper.

Fredhopper versions from 8.1

Start the initial reindex process. This will load the XML created in demo/data/fas-xml into FAS.

Unix:

bin/reindex-initial demo

Windows:

bin\reindex-initial.bat demo

Fredhopper versions before 8.1

Start the reindex process. This will load the XML created in demo/data/fas-xml into FAS.

Unix:

bin/reindex demo

Windows:

bin\reindex demo

Verify reindex process

After reindex or reindex-initial script has finished, open the Preview pages of the indexer instance: http://localhost:8180/preview, and open the Business Manager using the link at the top. The default username is admin, password admin.

The host and port will differ depending on the config/topology.txt configuration of the indexer instance.

You should see a page like

On Unix/Linux systems an additional script is available to show the number of items loaded: $ bin/chk-up demo Querydate Instancename Roundtriptime(ms) Host Port #Items Comments INFO yyyy-mm-ddThh:mm:ss demo 309 localhost 8180 17799 INFO yyyy-mm-ddThh:mm:ss live1 133 localhost 9180 0 INFO yyyy-mm-ddThh:mm:ss live2 130 localhost 10180 0 The last number shows the number of items in the instance. Note that the bin/reindex has only loaded the items in the indexer instance.

Step 8: Publish data and configuration to live machine

Execute these steps to bring the demo to all live qservers configured in config/topology.txt:

1. File the configuration by clicking File in Business Manager, and then File again. This will open the Publication Manager.
2. Publish the configuration by clicking Approve, and finally, click Publish
3. For FAS 8 users there is an extra manual step needed for each newly added live node, copy the following directory from the indexer to the live node:
   1. from: indexer: <HOME>/fredhopper/data/instances/<INSTANCE>/config/backup/<LATEST_TIMESTAMP>/*
   2. to: live: <HOME>/fredhopper/data/instances\<INSTANCE>/config
4. Push the data to live (this will establish a permanent connection between the indexer and the live site, allowing for publishes in the future to be pushed to the live site):
   Unix:

   bin/fresh-index-to-live demo

   Windows:

   bin\fresh-index-to-live demo

5. Open in a web browser http://<host>:<port>/preview for each live instance

On Unix/Linux systems you can use bin/chk-up demo now as well: $ bin/chk-up demo Querydate Instancename Roundtriptime(ms) Host Port #Items Comments INFO yyyy-mm-ddThh:mm:ss demo 309 localhost 8180 17799 INFO yyyy-mm-ddThh:mm:ss live1 133 localhost 9180 17799 INFO yyyy-mm-ddThh:mm:ss live2 130 localhost 10180 17799 After the bin/fresh-index-to-live has finished the items are now also loaded into the live1 and live2 instances.