title	sidebar_label	description
Quick start	Quick start	Get started with Onepanel on AKS, EKS or GKE

import Tabs from '@theme/Tabs'; import TabItem from '@theme/TabItem';

It's easy to get started with Onepanel. First, you install the CLI (opctl) and then using opctl, you generate a params.yaml file and update it to configure your deployment. Once complete, you can access your deployment from any browser using your Kubernetes authentication token. Finally, you can run a Workflow or create a Workspace.

:::important The steps in the Quick start allow you to quickly setup a Onepanel cluster for testing. To setup a production cluster with TLS and auto scaling enabled see instructions for your cloud provider.

You can also add components at later time to make this cluster production ready. :::

Before getting started, take a look at concepts to understand the different components in Onepanel.

Step 0: Create a Kubernetes cluster

Next, create a Kubernetes cluster in one of the following cloud providers:

:::note Make sure Azure CLI (az) is installed before proceeding. :::

Run this az command to create a bare minimum cluster with 2 Standard_D4s_v3 nodes:

az aks create --resource-group <resource-group> --name <cluster-name> --location <region> \
    --node-count 2 \
    --node-vm-size Standard_D4s_v3 \
    --node-osdisk-size 100 \
    --min-count 2 \
    --max-count 2 \
    --enable-cluster-autoscaler \
    --network-plugin azure \
    --network-policy azure \
    --enable-addons monitoring \
    --generate-ssh-keys

You can then get access credentials by running:

az aks get-credentials --resource-group <resource-group> --name <cluster-name> --admin

:::note Make sure Amazon EKS CLI (eksctl) is installed before proceeding. :::

Run this eksctl commands to create a bare minimum cluster with 2 m5.xlarge nodes: <Tabs groupId="operating-system" defaultValue="linux" values={[ { label: 'Linux', value: 'linux', }, { label: 'Windows', value: 'windows', }, ] }>

eksctl create cluster --name=<cluster-name> --region <region> \
    --nodes 2  \
    --node-type m5.xlarge \
    --node-volume-size 100 \
    --nodes-min 2 \
    --nodes-max 2 \
    --asg-access \
    --managed \
    --ssh-access

eksctl create cluster --name=<cluster-name> --region <region>^
    --nodes 2^
    --node-type m5.xlarge^
    --node-volume-size 100^
    --nodes-min 2^
    --nodes-max 2^
    --asg-access^
    --managed^
    --ssh-access

The eksctl command above will automatically retrieve your cluster's access credentials but you can also get them by running:

eksctl utils write-kubeconfig --cluster=<cluster-name> --region <region>

:::note Make sure Google Cloud SDK (gcloud) is installed before proceeding. :::

Run this gcloud command to create a bare minimum cluster with 2 n1-standard-4 nodes:

gcloud container --project <project-name> clusters create <cluster-name> --zone <zone> \
    --num-nodes 2 \
    --machine-type n1-standard-4 \
    --disk-size 100 \
    --min-nodes 0 \
    --max-nodes 2 \
    --enable-autoscaling \
    --enable-network-policy \
    --enable-stackdriver-kubernetes \
    --addons HorizontalPodAutoscaling,HttpLoadBalancing

The command above will automatically retrieve your cluster's access credentials but you can also get them by running:

gcloud container clusters get-credentials <cluster-name> --zone <zone>

:::note Make sure Minikube (minikube) is installed before proceeding. :::

Run the following minikube command to create a cluster: <Tabs groupId="operating-systems" defaultValue="linux" values={[ { label: 'Linux', value: 'linux', }, { label: 'macOS', value: 'macos', }, { label: 'Windows', value: 'windows', }, ] }>

minikube start --driver=virtualbox --memory '8gb' --cpus=4 --disk-size '40g' \
    --extra-config=apiserver.service-account-signing-key-file=/var/lib/minikube/certs/sa.key \
    --extra-config=apiserver.service-account-key-file=/var/lib/minikube/certs/sa.pub \
    --extra-config=apiserver.service-account-issuer=api \
    --extra-config=apiserver.service-account-api-audiences=api,nats \
    --extra-config=apiserver.authorization-mode=Node,RBAC

minikube start --driver=virtualbox --memory '8gb' --cpus=4 --disk-size '40g' \
    --extra-config=apiserver.service-account-signing-key-file=/var/lib/minikube/certs/sa.key \
    --extra-config=apiserver.service-account-key-file=/var/lib/minikube/certs/sa.pub \
    --extra-config=apiserver.service-account-issuer=api \
    --extra-config=apiserver.service-account-api-audiences=api,nats \
    --extra-config=apiserver.authorization-mode=Node,RBAC

minikube start --driver=virtualbox --memory="8gb" --cpus=4 --disk-size="40gb"^
    --extra-config=apiserver.service-account-signing-key-file=/var/lib/minikube/certs/sa.key^
    --extra-config=apiserver.service-account-key-file=/var/lib/minikube/certs/sa.pub^
    --extra-config=apiserver.service-account-issuer=api^
    --extra-config=apiserver.service-account-api-audiences=api,nats^
    --extra-config=apiserver.authorization-mode=Node,RBAC

Your kubectl context will be automatically updated once minikube finishes starting.

First, install Multipass for your operating system:

:::info Instructions See Installing on Linux :::

:::info Instructions See Installing on macOS :::

:::info Instructions See Installing on Windows :::

Run the following multipass command to launch Multipass:

multipass launch --name microk8s-vm --mem 8G --disk 40G --cpus 4 # 8G ram recommended for Istio

Multipass creates a virtual machine (VM). Inside that VM, we will create a Kubernetes cluster with microk8s.

Run a shell into your VM:

multipass shell microk8s-vm

Install the MicroK8s snap and configure the network:

sudo snap install microk8s --classic --channel=1.18/stable
sudo iptables -P FORWARD ACCEPT

You will also need to add ubuntu user to microk8s group as follows:

sudo usermod -a -G microk8s ubuntu

# Re-enter bash session for group changes
exit
multipass shell microk8s-vm

Then, enable the following required add-ons:

sudo microk8s.enable storage dns rbac dashboard

Enable TokenRequest feature (required by Istio) by passing in extra argument to kube-apiserver.

nano /var/snap/microk8s/current/args/kube-apiserver

Add the lines:

--service-account-signing-key-file=${SNAP_DATA}/certs/serviceaccount.key
--service-account-key-file=${SNAP_DATA}/certs/serviceaccount.key
--service-account-issuer=api
--service-account-api-audiences=api,nats

Make sure this line is set to these values:

--authorization-mode=RBAC,Node

Save your changes. Execute to make changes take effect

sudo systemctl restart snap.microk8s.daemon-apiserver

Exit out of the VM for the next steps.

exit

Next, get the kubeconfig by running

   multipass exec microk8s-vm -- /snap/bin/microk8s.config > kubeconfig

Step 1: Install Onepanel

Download the latest opctl for your operating system from our release page.

# Download the binary
curl -sLO https://github.com/onepanelio/core/releases/latest/download/opctl-linux-amd64

# Make binary executable
chmod +x opctl-linux-amd64

# Move binary to path
mv ./opctl-linux-amd64 /usr/local/bin/opctl

# Test installation
opctl version

# Download the binary
curl -sLO https://github.com/onepanelio/core/releases/latest/download/opctl-macos-amd64

# Make binary executable
chmod +x opctl-macos-amd64

# Move binary to path
mv ./opctl-macos-amd64 /usr/local/bin/opctl

# Test installation
opctl version

:::info Download the attached executable, rename it to opctl and move it to a folder that is in your PATH environment variable. :::

Run the following command to initialize a params.yaml template for your provider:

opctl init --provider aks

:::note If you have GPU nodes, you need to set the --gpu-device-plugins flag. Valid values are nvidia and amd or a comma separated combination of both nvidia,amd. :::

opctl init --provider eks

:::note If you have GPU nodes, you need to set the --gpu-device-plugins flag. Valid values are nvidia and amd or a comma separated combination of both nvidia,amd. :::

opctl init --provider gke

opctl init --provider minikube --enable-metallb

:::note metallb is used as a loadbalancer :::

:::note If you have GPU nodes, you need to set the --gpu-device-plugins flag. Valid values are nvidia and amd or a comma separated combination of both nvidia,amd. :::

opctl init --provider microk8s --enable-metallb

:::note metallb is used as a loadbalancer :::

:::note If you have GPU nodes, you need to set the --gpu-device-plugins flag. Valid values are nvidia and amd or a comma separated combination of both nvidia,amd. :::

:::important You can use different object storage than S3 by setting the --artifact-repository-provider flag.

See CLI overview for additional flags that that may apply. :::

Populate params.yaml by following the instructions in the template, you can also refer to configuration file sections for more detailed information.
Finally, run the following command to deploy Onepanel to your cluster:

opctl apply

opctl apply

opctl apply

opctl apply

KUBECONFIG=./kubeconfig opctl apply

:::note If you do not have a loadbalancer setup, you can use metallb

opctl init --provider microk8s --enable-metallb

:::

:::note If the command completes but it indicates that your cluster is not ready, you can check status again by running opctl app status. If you're still seeing issues, visit our Troubleshooting page. :::

Once the deployment completes, the CLI will display the host name and wildcard domain you need to use to setup your DNS. You can also get this information again by running:

opctl app status

Create the appropriate DNS record in your DNS provider based on the instructions above.

:::tip If you don't have a domain name handy or you're waiting for your DNS record to propogate, you can set up a hosts file to quickly test the deployment. :::

Wait a few minutes and check the URL mentioned in the instructions above. Your applications should load with a screen prompting you to enter a token.

:::note If the application is not loading, visit our Troubleshooting page for some steps that can help resolve most issues. If you are still having issues, join our Slack community or open an issue in GitHub. :::

Depending on your provider, you may need to take the following steps to access the application:

:::info No additional steps required. :::

To get access to this new cluster via browser, we need to carry out extra steps.

Example request flow

Execute these steps in the host machine.

multipass list

you should see something like this

Name                    State             IPv4             Image
microk8s-vm             Running           10.174.163.50    Ubuntu 18.04 LTS

Grab the IP address, 10.174.163.50 in this case.

Add an entry to your hosts file to point to the fqdn you setup in params.yaml For example, if our fqdn was set up like this

# The Fully Qualified Domain (FQDN) where Onepanel will be hosted.
# If `domain` above is set to example.com or sub.example.com, then your FQDN could be: app.example.com or app.sub.example.com respectively
fqdn: app.alex.xyz

Then we add this to /etc/hosts

10.174.163.50 app.alex.xyz

Adding this entry means the host browser will try to access the multipass vm we setup for microk8s.

Next, enter into the multipass VM

multipass shell microk8s-vm

:::note Execute inside the multipass VM

microk8s.kubectl get services -n istio-system

:::

NAME                   TYPE           CLUSTER-IP       EXTERNAL-IP   PORT(S)                                                                                                                      AGE
istio-ingressgateway   LoadBalancer   10.152.183.166   10.1.31.0     15020:31979/TCP,80:31394/TCP,443:30038/TCP,15029:32204/TCP,15030:32688/TCP,15031:31420/TCP,15032:30575/TCP,15443:30386/TCP   3d3h

Inside the multipass VM, add an entry to the /etc/hosts file using the EXTERNAL-IP from above.

sudo nano /etc/hosts

10.1.31.0 app.alex.xyz

Once you have entered and saved the host change, verify the onepanel website is running.

curl app.alex.xyz # This is the fqdn entry from params.yaml

Example output.

  <!doctype html>
  <html lang="en">
  <head>
    <meta charset="utf-8">
    <title>Onepanel</title>
    <base href="/">
    <meta name="viewport" content="width=device-width, initial-scale=1">
    <link rel="icon" type="image/png" sizes="32x32" href="assets/icon/favicon.png">
    <link rel="icon" type="image/png" sizes="96x96" href="assets/icon/favicon.png">
    <link rel="icon" type="image/png" sizes="16x16" href="assets/icon/favicon.png">
  <link rel="stylesheet" href="styles.9b8cd86ace5a9057a37e.css"></head>
  <body>
    <app-root></app-root>
  <script src="runtime-es2015.edb2fcf2778e7bf1d426.js" type="module"></script><script src="runtime-es5.edb2fcf2778e7bf1d426.js" nomodule defer></script><script src="polyfills-es5.6696c533341b95a3d617.js" nomodule defer></script><script src="polyfills-es2015.2987770fde9daa1d8a2e.js" type="module"></script><script src="main-es2015.b17adb3826cd9f5e4a29.js" type="module"></script><script src="main-es5.b17adb3826cd9f5e4a29.js" nomodule defer></script></body>
  </html>

You can debug the request with curl -vvv app.alex.xyz

We need a listener running on port 80. That listener should direct traffic to the istio gateway.

sudo apt install socat
sudo socat TCP-LISTEN:80,fork TCP:app.alex.xyz:80

This will run actively in the current terminal prompt.

Now, go back to your host machine, open your internet browser and go to: app.alex.xyz.

You should see the website load up.

Use the following command to get your auth token to log into Onepanel:

opctl auth token

Step 2: Launch CVAT for annotation

You can use the built-in CVAT to annotate visual data (images, videos). To launch CVAT, click on Workspaces, and you will find Create Workspace button as shown below.

Now, select CVAT, machine type, you can also modify other parameters such as disk size. Once done, click on CREATE AND RUN. This will launch a new CVAT workspace. Please note that minimum RAM requirement for CVAT is 16GB.

Once in CVAT, you can annotate imges with bounding boxes, polygons, polylines, points, etc.

Step 3: Train a model on annotated data

Onepanel CE allows you to train a deep learning model on annotated data in CVAT in a matter of a few clicks.

You can directly train object detection or segmentation model from CVAT which you can use later to pre-annotate new data. This dramatically reduces time it takes to train new models and use them to pre-annotate new data. Moreover, apart from default models that we provide, you easily add new models or even modify exising model workflows.

To train a new model, first make sure you have some annotated data in a task. Then, click Actions or Menu button for a CVAT task on which you want to train a model. Click on Execute training workflow. You will see following pop up window.

Here, you can select model to train. In this case, it's MaskRCNN. Most of the parameters visible above are related to the model (MaskRCNN). Number of classes should be equal to number of classes you have in that CVAT task + 1 for background. However, it is important to note that frames and corresponding annotation will be automatically dumped onto the cloud object storage (path: <your-namespace>/annotation-dump/<time-stamp>/) that you used while setting up Onepanel. Similary, output of a workflow (i.e trained model) will also be uploaded to cloud storage (path: <your-namespace>/workflow-data/<workflow-uid>/<time-stamp>/). Here, you can also select the checkpoint path from previously trained model. You can leave it empty if you don't have an appropriate, previously trained model.

Please note you can add your models as well. Once you select the model, you will see a list of parameter for that model (Onepanel Workflow) which you can modify or use default model. If your workflow does not have dump-format parameter set, then you will also have to select the Dump Format (i.e MS COCO or TFRecords) as well based on which workflow your model/workflow supports.

For more information about CVAT on Onepanel, please refer to our docs on CVAT.

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

quickstart.md

quickstart.md

Step 0: Create a Kubernetes cluster

Step 1: Install Onepanel

Step 2: Launch CVAT for annotation

Step 3: Train a model on annotated data

Files

quickstart.md

Latest commit

History

quickstart.md

File metadata and controls

Step 0: Create a Kubernetes cluster

Step 1: Install Onepanel

Step 2: Launch CVAT for annotation

Step 3: Train a model on annotated data