Description

This repository is created to demonstrate how we can deploy application on k8s cluster with Jenkins, Terraform and Helm. Intention is to showcase minimum viable product. Customizations are encouraged as required.

Inputs

Name	Description	Type	Default	Required
chart_version	Tensorflow chart version. More info: https://github.com/helm/charts/tree/master/stable/tensorflow-notebook	string	0.1.3	no
tensorboard_image_pullPolicy	Tensorboard image pullPolicy	string	Always	no
tensorboard_image_tag	TensorFlow Development image tag	string	latest	no
tensorboard_image_repository	TensorFlow Development image repo	string	tensorflow/tensorflow	no
service_type	The kubernetes service type	string	NodePort	no

Probable error you may face

1. Default user forbidden to perform tasks in kubernetes cluster.

   Error:

   I received below error when performing "terraform apply". Error: rpc error: code = Unknown desc = configmaps is forbidden: User "system:serviceaccount:kube-system:default" cannot list resource "configmaps" in API group "" in the namespace "kube-system"

   Reason:

   Default user in kube-system namespace does not have required permissions.

   Resolution:

   To make it simple, I added cluster-admin role to default id using yaml below.

    apiVersion: rbac.authorization.k8s.io/v1
    kind: ClusterRoleBinding
    metadata:
    name: default
    roleRef:
    apiGroup: rbac.authorization.k8s.io
    kind: ClusterRole
    name: cluster-admin
    subjects:
    - kind: ServiceAccount
    name: default
    namespace: kube-system

2. Deploy does not finish.

   Error:

   Terraform apply shows "..still creating.." till timeout. Although the release is created and working as expected.

   Reason:

   Helm 2 to Helm 3 migration. And terraform provider used is Helm 2 supportive. Terraform was creating helm 2 release. Kubernetes cluster was trying to find helm 3 release and not getting it. As a result, terraform was thinking release is not yet created and never finishing.

   Resolution:

   Follow instructions mentioned in below pages. Converting releases and cleaning up configuration willbe required. https://github.com/helm/helm-2to3 https://helm.sh/blog/migrate-from-helm-v2-to-helm-v3/

3. Helm list does not show release.

   Error:

   helm list does not show release althogh release is deployed and pods are working.

   Reason:

   Same as point 2.

   Resolution:

   Same as point 2.

4. Cluster out of resource.

   Error:

   Terraform is trying to created pods, but pods are throwing below error. 2 Insufficient cpu, 2 Insufficient memory.

   Reason:

   Cluster has limited memory and cpu. By default, chart tries to craete pods with certain cpu and memory. Multiple of these may exhaust cluster.

   Resolution:

   Add resource limit.