P4: Custos Deployment Custos Deployment Step 1

Custos Deployment Step 1 - Rancher Setup

Back to Home

There was an issue in the setup for the custos deployment when Rancher was setup in Jetstream 2. The following steps are done in Jetstream 1 instances

Setup SSH on all the nodes

Cloudmon Setup

The cloudmon zip is available in the below path. This is to be downloaded and extracted.

https://airavata.slack.com/files/U030JR7JXDF/F03CA28HZ6J/cloudman.zip 

Modify sample.ini

Open the file in the path Inventory/sample.ini in the above extracted folder.

The modifications to be done are marked with # in the below file.

Remove these comments while executing.

[controllers]
js2-154-102.jetstream-cloud.org #Your host name

[agents]
js2-154-102.jetstream-cloud.org #Your host name

[rke_cluster:children]
controllers
agents

[all:vars]
ansible_ssh_port=22 
ansible_user='exouser' #Username of the vm on jetstream 1
ansible_user_password='LUKE AWN CRAM SHOE GIRL BOOT HERO EVIL OWLY WOK YEAH' #Passphrase to the vm on jetstream 1
ansible_ssh_extra_args='-o StrictHostKeyChecking=no' 
ansible_ssh_private_key_file=/Users/chowtagowtham/.ssh/id_rsa #Local machine private key that is configured on the vm

Creating Virtual Environment

Change the directory to the cloudman-boot folder in the extracted folder before running the following commands

python -m venv cust-dev
source cust-dev/bin/activate
pip install ansible

Run Ansible Playbook

ansible-playbook -i inventory/sample.ini  playbook.yml

We faced some issues on this step while running on Jetstream 2. Some pods kept failing to deploy. ingress-nginx was going into CrashLoopBackOff state. We were unable to delete the namespaces and restart the process as they were stuck on terminating state. We fixed this issue by referring to the known issues section provided by the team, Team Terra.

Once this playbook runs successfully, we have Helm, Rancher, Cloudmon setup completed

Rancher

On the Jetstream vm, run the following commands:-

Modify the hostname and email in the last command below

helm repo add rancher https://releases.rancher.com/server-charts/stable
kubectl create namespace cattle-system
helm repo update
helm install -n cattle-system rancher rancher/rancher --set hostname=custos-m.raghunadham.com --set ingress.tls.source=letsEncrypt --set letsEncrypt.email="[email protected]" --set letsEncrypt.environment="production" --set letsEncrypt.ingress.class=nginx --version 2.5.5

It is successfully installed if the following is displayed.

Wait for the cluster to be started.

Checking status

kubectl get all --all-namespaces
kubectl get services  -n ingress-nginx

The output of the second command is as follows:-

From the above output, we have to take the port associated with 443 (https) and append it to https://domainname:port to access on the browser. Our url would be https://custos-m.raghunadham.com:32261/

The output of the above command should show all the pods in running or completed state. The deployments and services should also be running

Create password. Username is password. The url and username are prompted to set the password.

References

Thanks to Team Terra