We have been asked on any number of occasions whether Qlik Replicate can be deployed in a Kubernetes environment. The short answer is "yes". This repository contains the scripts and configuration YAML files needed to stand up a demonstration environment for you to experiment with.
Qlik fully supports the deployment of Qlik Replicate in a Docker image. However, Qlik does not currently maintain a publicly accessible repository where you can download an image. You will need to provide your own.
Note: while Docker is fully supported by Qlik and Qlik Replicate can be deployed in a Kubernetes cluster, Kubernetes deployments are not formally supported by Qlik at the present time. Consequently, Qlik's support team will likely not be able to assist with issues specific to the Kubernetes environment as they lack the necessary expertise.
An example Dockerfile and supporting scripts can be found in the docker subdirectory. You
very likely will want to customize this for the specific requirements of your organization,
particularly with regard to the database drivers that you choose to install.
If you don't have one already, you will need to build a Qlik Replicate docker image to work with.
- Create Kubernetes secrets to hold SSL certs, Replicate password, and your Replicate license file.
- Use a Kubernetes "deployment" for production.
- The replication factor for your Replicate deployment should be set to 1. Replicate does not support sharing the data directory in active/active fashion across multiple instances of Replicate.
- The Replicate "data directory" should be located on storage that can be shared with all nodes in your cluster unless you are going to lock Replicate to a specific node.
The data directory is used heavily by Qlik Replicate. In high volume use cases performance can be impacted if access to this shared storage is suboptimal.