Removing a Rim Node, not to be a Failover candidate in the cluster

Removing a Rim Node, not to be a Failover candidate can be done in two steps. Ether removing the Rim node from the Rim server itself and install it again without enabling the "Candidate failover" option, 
or by changing the configuration in the repository database. This article explains how to change the option in the repository database.

This article explains how to do it in the repository.

Starting with Qlik Sense June 2017 you can set a RIM node to be a failover candidate for the Central node. That fail over option can be changed in the repository database.
You'll need to download pgAdmin separately as it no longer comes bundled with Qlik Sense June 2017 (Installing and Configuring pgAdmin 4 for Qlik Sense June 2017 Release and up), or you can use another admin tool of your own choice.

Warning: Modifying the repository manually is not supported and need to be done with caution and prefer to take a backup of the database before changing anything in the repository database.

 

1. Stop all services except the  Qlik Sense Repository Database.

2. Once you've connected to the database, navigate to ServiceNodeConfiguration

3. Right-click and select "View/Edit Data > All Rows"

4. Scroll to the end of the table, and double-click the "FailoverCandidate boolean" value

5. Edit that value to false, and click "Save"

The option "Failover candidate" option has been disable for the Rim2 Node.