Quick Start
This quickstart will:
- Guide you on getting the tools necessary to interact with the NRP installed
- Configure your Kubernetes client to communicate with the NRP cluster
- Show the commands to query the NRP cluster to view running processes
- Point you to further resources to submit jobs and learn Kubernetes
Please Note
- Containers are stateless. ALL your data WILL BE GONE FOREVER when container restarts, unless you store in a persistent volume.
- Container restart is normal in k8s cluster. Expect it.
- NEVER FORCE DELETE PODS
- Users running a Job with "sleep" command or equivalent (any command that never ends by itself) will be banned from using the cluster.
Step by Step
- Install the kubectl tool
-
Login to NRP Nautilus portal and click the Get Config link on top right corner of the page to get your configuration file.
Same Login
Make sure you use the same provider and account you used to login to the portal. If you use the different provider with the same email, your account in the system will still be different and your namespaces membership will change.
-
Save the file as config and put the file in your
/.kube folder. This folder may not exists on your machine, to create it execute:
-
Any cluster admin can promote you to admin (ask in Matrix, provide your affiliation and short project description), or any admin can promote you to user by adding to one of namespaces (contact the admin directly). Make sure you are promoted from guest to user (by joining an existing namespace) or to admin by getting a confirmation from either admin or cluster admin.
-
If you've become an admin, you can start creating your own namespaces at this time by going to the Namespaces section in the portal. Also you can add other users on the same page after they've logged in to the portal. If you're joining an existing namespace as a user, make sure your namespace's admin added you to it. To verify it go to Namespaces link (while logged on the portal).
-
Test kubectl can connect to the cluster using a command line tool:
It's possible there are no pods in your namespace yet. If you've got
No resources found.
, this indicates your namespace is empty and you can start running in it. -
To learn more about kubernetes you can look at our tutorial.
These resources might be helpful:
- kubectl tool overview and cheatsheet
- kubernetes basics from kubernetes project
- tutorials
Please note that not all examples will work in our cluster because of security policies. You are limited to see what's happening in your own namespace, and nobody else can see your running pods.
-
MANDATORY: read the Policies page
-
Proceed to creating your first ML job in kubernetes
-
You might want to try one of these GUI tools for Kubernetes:
Both will use your config file in default location to get access to the cluster.