jelastic-jps / kubernetes

Kubernetes Clusters - Managed and Unmanaged Cloud Hosting

Kubernetes Clusters

The clusters can be automatically installed on top of Jelastic PaaS in two ways:

• Clean cluster with pre-deployed HelloWorld example
• Custom helm or stack deployment via shell command

Also, there are different topologies available out-of-box:

• Development - one control-plane (1) and one scalable worker (1+)
• Production - multi control-plane (3) with API balancers (2+) and scalable workers (2+)

The number of worker nodes can be changed after the initial installation is completed. All the newly added nodes are automatically connected to the cluster.

Cloud Native Applications

Easily containerize and migrate existing applications, run hyper scalable microservices and keep them resilient to failures, get extra savings due to more efficient resource utilization, implement CI/CD automation and develop at a higher level of speed in shorter release cycles.

Jelastic PaaS functionality allows to provision the clusters across multiple clouds and on-premises with no vendor lock-in, automatically scale them vertically and horizontally, start from one instance and grow up to thousands, manage the workloads via intuitive UI, as well as automate the DevOps processes with open API and Cloud Scripting.

Related Articles and Materials

• Managed Kubernetes Hosting with Multi-Cloud Availability
• Kubernetes Cluster Setup with Automated Scaling and Pay-per-Use Pricing
• Scaling Kubernetes on Application and Infrastructure Levels
• Kubernetes Cluster Overview
• Kubernetes Cluster: Package Installation
• Kubernetes Cluster: Versions & Change Logs

Demo and Trial

For testing in public cloud please sign up at one of Jelastic Cloud Providers with Kubernetes support. For testing in multi-cloud, hybrid cloud or private cloud setups please send us request.

Note: The installation of this clustered solution is available only for billing customers

Managed Hosting Business

Interested in building managed hosting business offering Kubernetes services? Automated installation, scaling, updates and integrated billing simplify the management and support of customers requiring little to no involvement. Start offering managed Kubernetes hosting to your users with minimal investment of time and efforts.

kubernetes's Issues

Impossible to easily connect to the cluster with Lens IDE

Hi,

Some time ago, it was possible to easily connect to the deployed k8s cluster with Lens IDE. That tool is very handy to develop stuff on k8s. Here's the thing: if I follow these instructions, then I can happily interact with the cluster by means of kubectl cli. That's a good start. However, I cannot connect to the cluster through Lens. It says

Authentication proxy started
Refreshing connection status ...
Invalid credentials

I guess that Lens does not support user authentication with the token. Instead of following these instructions, I did the following:

1. install the OpenVPN Access Server from the Jelastic Marketplace
2. install OpenVPN client on my computer
3. connect my VPN client to my VPN server
4. in my k8s cluster, connect to my master node and display the /root/.kube/config file
5. take over the remote kube config and copy / paste it to my local kube config
6. replace the k8sm.env-8930174839.hidora.com hostname with the k8sm node's ip address

This amounts to take the configuration obtained through these instructions and

• replace section

- cluster:
    server: https://env-9878152.jelastic.com/api
  name: jelastic-test

with

- cluster:
    certificate-authority-data: <data from remote cluster>
    server: https://<private-k8sm-node-ip>:6443
  name: jelastic-test

• replace section

- name: jelastic-test
  user:
    token: <user-access-token>

with

- name: jelastic-test
  user:
    client-certificate-data: <client-certificate-data>
    client-key-data: <client-key-data>

With that in place, I can happily connect to the cluster with Lens. Is there a way to achieve that without VPN client? With Azure, this step is made very easy with their cli. You get the "credentials" with the azure cli and your kube config is updated with exactly the above working dataset (cluster certificate-authority-data, user client-certificate-data and user client-key-data). Would something like that be also possible on Jelastic?

The access token is not displayed correctly in confirmation emails

After installing version v1.25.4, here's how the confirmation email looks like:

There seems to be an issue with the rendering of the access token, as you can see.

yaml deployment script seems supported in manifest but not in settings

In your manifest, here, you seem to accept a yaml script, however in your settings, you don't propose the option here.

No letsencrypt addon available from the dashboard with version v1.16.6

I've tried today to install following manifest

https://github.com/jelastic-jps/kubernetes/blob/v1.16.6/manifest.jps

Though it works, i.e. it can be installed (it was not the case last week), I see no let's encrypt addon when I click the addon button on the worker nodes:

Is that intended? Is the only possible way to install that addon through the manifest (which I haven't tried yet)?

Kubernetes API Not Reachable

I tried the install Kubernetes multiple times over the last few weeks on Jelastic. I was not able to get the default installation running with the API reachable. I expect to use the main URL with port 6443.
I always had to create a custom endpoint to get access to the API. But then I receive error messages like:

Get https://registry-1.docker.io/v2/: x509: certificate is valid for *.jpc.infomaniak.com, jpc.infomaniak.com, not registry-1.docker.io

I already contacted the Infomaniak (hosting provider) team and they contacted you (the Jelastic team) but they haven't found the problem.

node-problem-detector pods stuck in pending state

after successfully installing kubernetes cluster, node problem detector pods fail to start :

  Type     Reason     Age                From               Message
  ----     ------     ----               ----               -------
  Normal   Scheduled  40s                default-scheduler  Successfully assigned default/node-problem-detector-99zjl to node38951-env-7450545.clicketcloud.com
  Normal   Pulling    26s (x2 over 40s)  kubelet            Pulling image "k8s.gcr.io/node-problem-detector/node-problem-detector:v0.8.11"
  Warning  Failed     26s (x2 over 40s)  kubelet            Failed to pull image "k8s.gcr.io/node-problem-detector/node-problem-detector:v0.8.11": rpc error: code = NotFound desc = failed to pull and unpack image "k8s.gcr.io/node-problem-detector/node-problem-detector:v0.8.11": failed to resolve reference "k8s.gcr.io/node-problem-detector/node-problem-detector:v0.8.11": k8s.gcr.io/node-problem-detector/node-problem-detector:v0.8.11: not found
  Warning  Failed     26s (x2 over 40s)  kubelet            Error: ErrImagePull
  Normal   BackOff    12s (x2 over 39s)  kubelet            Back-off pulling image "k8s.gcr.io/node-problem-detector/node-problem-detector:v0.8.11"
  Warning  Failed     12s (x2 over 39s)  kubelet            Error: ImagePullBackOff

Is this related to kubernetes/node-problem-detector#692 ?

Unable to deploy using Garden.io: 'error upgrading connection'

I have a Jelastic instance at MassiveGrid (trial still for now). I'm trying to install OpenFaas (as an example) using Garden.io. However, I'm having a strange issue where (a fresh) Kubernetes complains that a port cannot be forwarded. Is there a possibility that you could check out what goes wrong in the cluster? I do not have this problem on GAE.

I confined the issue into this repository: https://github.com/Tails/jelastic-kubernetes-garden-openfaas