How to develop the ESI API in an OpenShift AI Workbench

Pull down Kafka secrets to your OpenShift AI Workbench

install -d -o $USER -g $USER -m 771 /opt/app-root/src/kafka/truststore/
install -d -o $USER -g $USER -m 771 /opt/app-root/src/kafka/keystore/
oc extract secret/esi --to=/opt/app-root/src/kafka/keystore/ --keys=user.crt --confirm
oc extract secret/esi --to=/opt/app-root/src/kafka/keystore/ --keys=user.key --confirm
oc extract secret/esi --to=/opt/app-root/src/kafka/keystore/ --keys=user.p12 --confirm
oc extract secret/kafka-cluster-ca-cert --to=/opt/app-root/src/kafka/truststore/ --keys=ca.crt --confirm
oc extract secret/kafka-cluster-ca-cert --to=/opt/app-root/src/kafka/truststore/ --keys=ca.p12 --confirm

How to connect to ESI API

To establish a connection, you'll need to:

1. put the clouds.yaml file in ~/.config/openstack/clouds.yaml. You can refer to the example_clouds.yaml for a sample clouds.yaml.

2. set the environment variable CLOUD_NAME. For the provided example_clouds.yaml, the CLOUD_NAME should be set to "openstack".

How to run the application in an OpenShift AI workbench

Before running the application, update the following environment variables to use unique values starting with "esi-" to avoid conflicts between workbenches:

KAFKA_GROUP_FULFILL
KAFKA_TOPIC_FULFILL_OFFER
KAFKA_TOPIC_ORDER_LOOP
KAFKA_TOPIC_ORDER_STATUS

To run the application:

git clone https://github.com/CCI-MOC/demo-esi-api.git ~/demo-esi-api
cd ~/demo-esi-api
pip install -r requirements.txt
python app.py

In a separate terminal, test the request:

curl -X GET http://localhost:8081/api/v1/nodes/list
curl -X GET http://localhost:8081/api/v1/networks/list
curl -X GET http://localhost:8081/api/v1/offers/list
# Attach the network to the requested nodes:
curl -X POST http://localhost:8081/api/v1/baremetal-order/fulfill \
  -H "Content-Type: application/json" \
  -d '{"order_id": "123_xyz",
       "network_id": "provisioning",
       "nodes": [{"resource_class": "fc430", "number": 2}],
       "create_floating_ip": "True"
  }'
# Or provision the nodes with metalsmith
curl -X POST http://localhost:8081/api/v1/baremetal-order/fulfill \
  -H "Content-Type: application/json" \
  -d '{"order_id": "123_xyz",
       "network_id": "network-uitest",
       "nodes": [{"resource_class": "fc430", "number": 1}],
       "create_floating_ip": "True",
       "image": "centos-image",
       "ssh_keys": [<ssh_key_string1>, <ssh_key_string2>...]
  }'

How to run the application as a Podman container

Install the prerequiste packages for buildah and podman

pkcon install -y buildah
pkcon install -y podman

Build the container with podman

cd ~/.local/src/TLC
podman build -t nerc-images/demo-esi-api:computate-api .

Push the container up to quay.io

podman login quay.io
podman push nerc-images/demo-esi-api:computate-api quay.io/nerc-images/demo-esi-api:computate-api

Run the container for local development

podman run -v /opt/kafka:/opt/kafka nerc-images/demo-esi-api:computate-api