Since Kubernetes v1.19, AKS has shifted
. However, the workflow controller installed with the Kubeflow installation has
Dockerset as the default runtime. In order to make your pipelines work, you have to change the value to one of the options
k8sapi.This change has to be made by editing the
containerRuntimeExecutorproperty of the
ConfigMapcorresponding to the workflow controller. Run the following commands to first know what config map to change and then to edit it to reflect your new value.kubectl get configmap -n kubeflowkubectl edit configmap CONFIGMAP_NAME -n kubeflow# This opens up an editor that can be used to make the change.
dockerCLI with your ECR registry using the following command. Replace the capitalized words with your values.
dockercli with your ACR registry using the following command.
AWS_SESSION_TOKENenvironment variables in your EKS cluster. This can be done by extending the
zenmldocker/zenmlimage and adding these variables in a Dockerfile.
AZURE_STORAGE_ACCOUNT_NAMEand one of [
AZURE_STORAGE_SAS_TOKEN]. This can be done
setx AZURE_STORAGE_CONNECTION_STRING "<MY_CONNECTION_STRING>"in the Command Prompt.
mysqlclient and require SSL for the access for the user you created at the first step. You will also need the AWS server certificate associated with the AWS region where the RDS instance is running, as covered in the AWS documentation. Download the certificate as a file and have it ready for use during the ZenML stack registration.
mysqlflavor metadata-store. You will also need a secrets manager component to store the MySQL credentials and a secret (named
mysql_secretin the example) to be registered after the stack (step 5.).
mysqlmetadata store (not necessary if you used a