Fredhopper - Scaled Out Setup

Question

I'm struggling to find information on scaling out Fredhopper (Indexer / Query Servers). I created an AIO machine and I can see that its working perfectly with SDL Experience Optimisation 8.5

Our plan is to have 1 x Index Server, 2 x Query Servers and I have a few questions.

• Do you set a topology only on the Indexer Server or do you copy the same topology to each server?
• Is the below topology valid for Fredhopper?

liveindexer|bbeuind01|1|-|Indexer

query1|bbeuque01|1|liveindexer|Query1

query2|bbeuque02|1|liveindexer|Query2

• How do you install just a Indexer / Query Server

Any help would be greatly appreciated as I'm a little confused by the documentation.

Answer 1

Used the following topology -

liveindexer|bbeuind01|1|-|Indexer and Preview environment for demo setup query1|bbeuque01|1|liveindexer|Live Query Server for demo setup query2|bbeuque02|1|liveindexer|Live Query Server for demo setup

When running bin\setup-cluster, this communicates with the deployment agents on each of the servers (specified in column 2 in your topology). I had to open up firewall ports for this to work!

You can view the log of Setup-Cluster via FREDHOPPERINSTALLPATH\data\instances\deployment\log\kitchen.log

You should see something like this -

INFO  25-07 14:16:38,395 - Create Instance Live - Starting...
INFO  25-07 14:16:43,625 - Create Instance Live - Creating instance 'query2' for installation 'FAS' with preset name '1' succeeded.
INFO  25-07 14:16:43,639 - Copied 500/-1 bytes
INFO  25-07 14:16:43,713 - Create Instance Live - Adding role from definition file 'http://10.200.0.4:8177/static/live.xml' to instance 'query2' succeeded.
INFO  25-07 14:16:43,713 - Create Instance Live - Finished operation of Job Entry [Create Instance Live] (result=[true])
INFO  25-07 14:16:43,713 - Install single instance - Starting entry [Set replication source]
INFO  25-07 14:16:43,714 - Set replication source - Starting...
INFO  25-07 14:16:43,781 - Set replication source - Setting option 'syncclient_server' to value 'bbeuind01' on instance 'query2' succeeded.
INFO  25-07 14:16:43,793 - Set replication source - Setting option 'syncclient_serverport' to value '8100' on instance 'query2' succeeded.
INFO  25-07 14:16:43,793 - Set replication source - Finished operation of Job Entry [Set replication source] (result=[true])
INFO  25-07 14:16:43,794 - Install single instance - Starting entry [Success Live]
INFO  25-07 14:16:43,796 - Install single instance - Finished job entry [Success Live] (result=[true])
INFO  25-07 14:16:43,796 - Install single instance - Finished job entry [Set replication source] (result=[true])
INFO  25-07 14:16:43,796 - Install single instance - Finished job entry [Create Instance Live] (result=[true])
INFO  25-07 14:16:43,796 - Install single instance - Finished job entry [Allow only live nodes] (result=[true])
INFO  25-07 14:16:43,796 - Install single instance - Finished job entry [Cluster definition to variable] (result=[true])
INFO  25-07 14:16:43,798 - Install single instance [ensure success] - Starting entry [Success]
INFO  25-07 14:16:43,805 - Install single instance [ensure success] - Finished job entry [Success] (result=[true])
INFO  25-07 14:16:43,805 - Install single instance [ensure success] - Finished job entry [Install single instance] (result=[true])
INFO  25-07 14:16:43,805 - Install single instance [ensure success] - Finished job entry [Cluster definition to variable] (result=[true])
INFO  25-07 14:16:43,814 - Setup Cluster - Finished job entry [Install single instance] (result=[true])
INFO  25-07 14:16:43,814 - Setup Cluster - Finished job entry [Read Cluster Topology] (result=[true])
INFO  25-07 14:16:43,815 - Setup Cluster - Finished job entry [Set directories] (result=[true])
INFO  25-07 14:16:43,815 - Setup Cluster - Job execution finished
INFO  25-07 14:16:43,815 - Kitchen - Finished!
INFO  25-07 14:16:43,816 - Kitchen - Start=2017/07/25 14:16:32.045, Stop=2017/07/25 14:16:43.816
INFO  25-07 14:16:43,816 - Kitchen - Processing ended after 11 seconds.

If anyone sees any problems with my setup please let me know but seems to be working for me.

Answer 2

You need to set topology on each server and it is not the same. On FH Indexer Server in topology you set both Indexer instance and Query instances. On FH Query Servers in topology you set only Query instances.

Topology is not valid because you set preset number to be the same for all instances (all configured to 1) and this is not valid. FH uses that preset number (among other things) to configure ports for each instance and each instance on one server needs to have unique preset number.

I assume that on one FH server you want both Indexer and Query instances to be installed so for this server you can follow default FH installation (unzip fredhopper package, configure topology, run setup-cluster, ...). On other server where you want to have only Query you can skip setup-cluster eecution and after your deployment agent is up and running you can follow FH documentation for adding aditional Query instance.