MLRun CE installation notes

MLRun CE installation notes#

This page lists additional steps or configuration options you may need to follow for non-default MLRun CE installations.

In this section

Advanced chart configuration#

Configurable values are documented in the values.yaml, and the values.yaml of all sub charts. Override those in the normal methods.

See also the MLRun CE values file reference

Opt out of components#

The chart installs many components. You may not need them all in your deployment depending on your use cases. To opt out of some of the components, use the following helm values:

--set pipelines.enabled=false \
--set kube-prometheus-stack.enabled=false \
--set spark-operator.enabled=false \

Using NFS storage#

If you are using NFS storage in your Kubernetes cluster, add these flags to the chart deployment command:

  --set kube-prometheus-stack.grafana.securityContext.runAsUser=1000 
  --set kube-prometheus-stack.grafana.securityContext.runAsGroup=1000 
  --set kube-prometheus-stack.grafana.securityContext.fsGroup=1000 
  --set kube-prometheus-stack.grafana.securityContext.fsGroupChangePolicy=OnRootMismatch 
  --set kube-prometheus-stack.grafana.initChownData.enabled

Configuring the online feature store#

The MLRun Community Edition supports the online feature store. To enable it, you need to first deploy a Redis service that is accessible to your MLRun CE cluster. To deploy a Redis service, refer to the Redis documentation.

When you have a Redis service deployed, you can configure MLRun CE to use it by adding the following helm value configuration to your helm install command:

--set mlrun.api.extraEnvKeyValue.MLRUN_REDIS__URL=<redis-address>

Installing Spark Operator on non-mlrun namespace#

By default Spark Operator jobNamespaces is set to "mlrun" namespace. If you are installing Spark Operator on a different namespace you need to set the jobNamespaces value accordingly

--set spark-operator.jobNamespaces={your-namespace}