Redis usage and documentation in production server

In the latest version, 3.7, the upgrade notes mention a change to Redis. Redis is also mentioned in the Kubernetes section of 3.6 and 3.7. However, I don’t see any documentation on how to actually configure Redis in the Kubernetes cluster.

We’ve upgraded our clusters from 3.4 to 3.6 and then to 3.7 using Helm, but nowhere along that process did Redis get installed. I see in the charts that device_auth references Redis with some variables, but when I inspect my pod the relevant Redis environment variables are not even present.

What would Redis even be doing in the cluster? It appears to not be a prerequisite, but it is mentioned as a dependency. What would be the necessary modifications to get it working?

Hi @akin ,
the official helm chart by default set up a Redis Subchart. However, please note that the services are using Redis only in the Enterprise version.

Thanks for the clarification! That would make sense as we’re not using Enterprise.

It is definitely confusing as with a new deployment, the stateful sets mender-redis-master and mender-redis-replicas are created. This can cause issues where pods are limited (AWS EKS as an example).

The docs should probably note to disable it, or to have it disabled by default and note to enable for enterprise customers.

Hi @akin ,
makes sense to disable it by default on non-enterprises setup, thanks for the feedback.

In the meantime, please take a look at the Mender Helm doc: mender-helm/README.md at master · mendersoftware/mender-helm · GitHub

In particular, to disable the default redis subchart, you can set:

redis:
  disable: true
1 Like