### NooBaa Installation Output

Source: https://github.com/datashim-io/datashim/wiki/Noobaa

This is an example of the expected output after running the NooBaa installation script. It confirms the successful download and installation of NooBaa, creation of backing stores and bucket classes, and loading of data to an example bucket.

```bash
$ ./examples/noobaa/noobaa_install.sh 
Downloading NooBaa CLI...done
Installing NooBaa...
Noobaa pods ready!
Installed NooBaa
Creating Backing Store
Created Backing Store
Backingstore ready!
Delete Bucket Class
Delete Bucket Class
Creating Bucket Class
Created Bucket Class
done
Creating test OBC...done
Loading data to example bucket...job.batch/example-noobaa-data condition met
done

```

--------------------------------

### Install NooBaa

Source: https://github.com/datashim-io/datashim/wiki/Noobaa

Execute this script from the root directory of the repository to install NooBaa and load sample data. The installation process includes downloading the NooBaa CLI, installing NooBaa, creating a backing store, and setting up a bucket class.

```bash
./examples/noobaa/noobaa_install.sh
```

--------------------------------

### Local Kubernetes Cluster Setup

Source: https://github.com/datashim-io/datashim/blob/master/src/csi-driver-nfs/README.md

Command to start a local Kubernetes cluster with required configurations for the CSI driver.

```shell
ALLOW_PRIVILEGED=true FEATURE_GATES=CSIPersistentVolume=true,MountPropagation=true RUNTIME_CONFIG="storage.k8s.io/v1alpha1=true" LOG_LEVEL=5 hack/local-up-cluster.sh
```

--------------------------------

### Install Scheduler with Helm

Source: https://github.com/datashim-io/datashim/blob/master/examples/scheduling/README.md

Navigate to the scheduler-plugins directory and install the scheduler using Helm.

```bash
$ cd $DATASHIM_HOME/examples/scheduling/src/sigs.k8s.io/scheduler-plugins
$ helm install cacheaware manifests/install/charts/as-a-second-scheduler/
```

--------------------------------

### Start Minikube for Testing

Source: https://github.com/datashim-io/datashim/blob/master/docs/Ceph-Caching.md

Starts a Minikube instance with specified memory, CPU, and disk size. This is for testing purposes.

```bash
minikube start --memory='6G' --cpus=4 --disk-size='40g' --driver=virtualbox -p rooktest
```

--------------------------------

### Pod Status After Dataset Creation Example

Source: https://github.com/datashim-io/datashim/blob/master/src/dataset-operator/README.md

Example output showing pod statuses after a Dataset object has been created. Includes the newly created example-dataset-pod.

```text
NAME                               READY   STATUS    RESTARTS   AGE
example-dataset-pod                1/1     Running   0          20s
dataset-operator-644f8d854-dct95   1/1     Running   0          3m50s
```

--------------------------------

### Start Minikube Cluster

Source: https://github.com/datashim-io/datashim/blob/master/examples/scheduling/README.md

Starts a minikube cluster with specific Kubernetes version and driver. This is a prerequisite for deploying the scheduler plugin.

```bash
minikube start --nodes 2 -p scheduler-test --kubernetes-version=v1.21.0 --driver=virtualbox 
```

--------------------------------

### Operator Pod Status Example

Source: https://github.com/datashim-io/datashim/blob/master/src/dataset-operator/README.md

Example output showing the status of the dataset operator pod. Verifies the operator is running.

```text
NAME                               READY   STATUS    RESTARTS   AGE
dataset-operator-644f8d854-dct95   1/1     Running   0          15s
```

--------------------------------

### Install Rook Ceph Common Configuration

Source: https://github.com/datashim-io/datashim/wiki/Ceph-Caching

Applies the common configuration for Rook Ceph. This is a prerequisite for further Rook/Ceph setup.

```bash
kubectl create -f common.yaml
```

--------------------------------

### Install Datashim using Helm

Source: https://github.com/datashim-io/datashim/blob/master/docs/index.md

Installs Datashim into the 'dlf' namespace using Helm. Ensure the namespace is created if it does not exist.

```bash
helm install --namespace=dlf --create-namespace \
              datashim datashim/datashim-charts
```

--------------------------------

### Start Minikube Cluster

Source: https://github.com/datashim-io/datashim/wiki/Ceph-Caching

Initiates a Minikube cluster with specified memory, CPU, and disk size. Ensure Docker is available as the driver.

```bash
minikube start --memory='10G' --cpus=8 --disk-size='15g' --driver=docker
```

--------------------------------

### Ceph Cluster Configuration Example

Source: https://github.com/datashim-io/datashim/blob/master/docs/Ceph-Caching.md

Example configuration for Ceph cluster storage, specifying nodes and devices. Adjust 'storage.nodes' and 'devices' according to your setup.

```yaml
storage:
    useAllNodes: false
    useAllDevices: false
    nodes:
      - name: "minikube"
        devices: 
          - name: "sdb"
            config:
              storeType: bluestore
              osdsPerDevice: "1"
```

--------------------------------

### Deploy Example Nginx Application

Source: https://github.com/datashim-io/datashim/blob/master/src/csi-driver-nfs/README.md

Command to deploy an example Nginx application that uses the CSI NFS driver. Ensure NFS server and share information are updated in the YAML file.

```shell
kubectl -f examples/kubernetes/nginx.yaml create
```

--------------------------------

### Deploy Minio and Datashim Components

Source: https://github.com/datashim-io/datashim/wiki/Archive-based-Datasets

Install Minio and other Datashim components using make and kubectl. Ensure you are in the 'dlf' namespace.

```bash
make deployment
kubectl apply -n dlf -f examples/minio/
```

--------------------------------

### Install Datashim using Helm

Source: https://github.com/datashim-io/datashim/blob/master/README.md

Installs Datashim using a specified version from the Helm repository into the 'dlf' namespace, creating the namespace if it does not exist.

```bash
helm install --namespace=dlf --create-namespace datashim datashim/datashim-charts --version <version_string>
```

--------------------------------

### Kubernetes E2E Test Setup (Fedora/RHEL)

Source: https://github.com/datashim-io/datashim/blob/master/src/csi-driver-nfs/README.md

Commands to set up permissions for running Kubernetes End-to-End tests on Fedora/RHEL clusters.

```shell
sudo chown -R $USER:$USER /var/run/kubernetes/
sudo chown -R $USER:$USER /var/lib/kubelet
sudo chcon -R -t svirt_sandbox_file_t /var/lib/kubelet
```

--------------------------------

### Install Hive on Minikube

Source: https://github.com/datashim-io/datashim/blob/master/examples/hive/k8s/README.md

Execute the make command to install Hive on a Minikube cluster. Ensure Makefile variables like DATASET_NAMESPACE_OPERATOR, DOCKER_REGISTRY_COMPONENTS, and DOCKER_REGISTRY_SECRET are correctly set beforehand.

```bash
$ make minikube-install
```

--------------------------------

### Monitor Datashim Installation Status

Source: https://github.com/datashim-io/datashim/wiki/Archive-based-Datasets

Watch the status of pods in the 'dlf' namespace to confirm successful installation of all components.

```bash
watch kubectl get pods -n dlf
```

--------------------------------

### Create Example Dataset

Source: https://github.com/datashim-io/datashim/blob/master/docs/index.md

Create a Datashim Dataset with S3 (COS) configuration. Avoid hardcoding credentials in production; use Secrets instead.

```yaml
cat <<EOF | kubectl apply -f -
apiVersion: datashim.io/v1alpha1
kind: Dataset
metadata:
  name: example-dataset
spec:
  local:
    type: "COS"
    accessKeyID: "{AWS_ACCESS_KEY_ID}"
    secretAccessKey: "{AWS_SECRET_ACCESS_KEY}"
    endpoint: "{S3_SERVICE_URL}"
    bucket: "{BUCKET_NAME}"
    readonly: "true" #OPTIONAL, default is false
    region: "" #OPTIONAL
EOF
```

--------------------------------

### Start NFS Driver

Source: https://github.com/datashim-io/datashim/blob/master/src/csi-driver-nfs/README.md

Start the NFS driver with specified endpoint, node ID, and log level.

```shell
sudo ./_output/nfsplugin --endpoint tcp://127.0.0.1:10000 --nodeid CSINode -v=5
```

--------------------------------

### Install Intermediate Datashim Helm Release

Source: https://github.com/datashim-io/datashim/blob/master/README.md

Installs an intermediate (e.g., alpha, beta) Datashim release using the --devel flag and a specified version name.

```bash
helm install --namespace=dlf --create-namespace datashim datashim/datashim-charts --devel --version <version_name>
```

--------------------------------

### Test Hive Installation

Source: https://github.com/datashim-io/datashim/blob/master/examples/hive/k8s/README.md

Run the test-hive.sh script to verify the Hive installation. Modify the namespace and repository variables within the script as needed before execution.

```bash
$ ./test-hive.sh
```

--------------------------------

### Setup Ceph Cluster with CSI-LVM

Source: https://github.com/datashim-io/datashim/blob/master/docs/Ceph-Caching.md

Shell script to set up a Ceph cluster using CSI-LVM on specified nodes. It will uninstall previous installations if run without arguments.

```bash
cd plugins/ceph-cache-plugin/deploy/rook && \
./setup_ceph_cluster.sh nodename1 ...
```

--------------------------------

### Create Rook Ceph Common and Operator Resources

Source: https://github.com/datashim-io/datashim/blob/master/docs/Ceph-Caching.md

Installs the common configuration and operator for Rook Ceph. Ensure you are in the correct directory.

```bash
kubectl create -f common.yaml
kubectl create -f operator.yaml
```

--------------------------------

### Configure and Install Ceph Keys

Source: https://github.com/datashim-io/datashim/wiki/Ceph-Caching

Executes a script to set up Ceph registry credentials. Replace placeholders with your actual Docker registry details.

```bash
./keys-installation.sh
```

--------------------------------

### Create Datashim Namespace

Source: https://github.com/datashim-io/datashim/blob/master/README.md

Creates the 'dlf' namespace for Datashim installation if it doesn't exist.

```bash
kubectl create ns dlf
```

--------------------------------

### Create Storage Class

Source: https://github.com/datashim-io/datashim/blob/master/src/csi-s3/README.md

Create a storage class resource that defines how storage should be provisioned. This example assumes a storage class named 'csi-s3' is defined in storageclass.yaml.

```bash
kubectl create -f storageclass.yaml
```

--------------------------------

### Create S3 Dataset Example

Source: https://github.com/datashim-io/datashim/wiki/Ceph-Caching

Defines and creates an S3 dataset using a Kubernetes Dataset resource. Replace placeholders with your S3 credentials and endpoint.

```yaml
apiVersion: com.ie.ibm.hpsys/v1alpha1
kind: Dataset
metadata:
  name: example-dataset
spec:
  local:
    type: "COS"
    accessKeyID: "{AWS_ACCESS_KEY_ID}"
    secretAccessKey: "{AWS_SECRET_ACCESS_KEY}"
    endpoint: "{S3_SERVICE_URL}"
    bucket: "{BUCKET_NAME}"
    region: "" #it can be empty
```

--------------------------------

### Dataset with Automatic Bucket Provisioning

Source: https://github.com/datashim-io/datashim/blob/master/docs/user-guide.md

Configure a Dataset to have Datashim automatically create the referenced bucket. This simplifies setup by managing bucket lifecycle.

```yaml
apiVersion: datashim.io/v1alpha1
kind: Dataset
metadata:
  name: my-dataset
spec:
  local:
    provision: "true" # <----
    bucket: my-bucket
    endpoint: http://my-s3-endpoint
    secret-name: my-dataset-secret
    type: COS
```

--------------------------------

### Verify Deployment Status

Source: https://github.com/datashim-io/datashim/wiki/Installation

Check the status of deployed components to ensure successful installation. All components should be in a 'Running' state.

```bash
NAME                               READY   STATUS              RESTARTS   AGE
csi-attacher-nfsplugin-0           2/2     Running             0          78s
csi-attacher-s3-0                  1/1     Running             0          78s
csi-nodeplugin-nfsplugin-j4ljv     2/2     Running             0          78s
csi-provisioner-s3-0               1/1     Running             0          78s
csi-s3-2gwcs                       2/2     Running             0          79s
dataset-operator-76f795587-cljfm   1/1     Running             0          77s
generate-keys-q8s99                0/1     Completed           0          67s
```

--------------------------------

### Configure and Install Ceph Cluster

Source: https://github.com/datashim-io/datashim/wiki/Ceph-Caching

Applies the cluster configuration for Rook Ceph. Ensure the `YOUR_REGISTRY` placeholder is replaced.

```bash
kubectl create -f cluster.yaml
```

--------------------------------

### Successful Hive Installation Verification

Source: https://github.com/datashim-io/datashim/blob/master/examples/hive/k8s/README.md

Check for an 'HTTP/1.1 200 OK' response after running the test script. This indicates a successful installation and allows verification via the Hive landing page URL in a browser.

```text
HTTP/1.1 200 OK
```

--------------------------------

### Verify Rados Gateway Pod

Source: https://github.com/datashim-io/datashim/wiki/Ceph-Caching

Lists pods in the rook-ceph namespace to confirm the Rados Gateway pod is starting up after the caching plugin deployment.

```bash
rook-ceph-rgw-test-a-77f78b7b69-z5kp9              1/1     Running     0          4m43s
```

--------------------------------

### Build Model Container Image

Source: https://github.com/datashim-io/datashim/wiki/Kubeflow-Model-Storing-and-Serving-with-DLF

Builds a Docker image for the MNIST model and pushes it to a container registry. Ensure you have Docker installed and are logged into your registry.

```bash
cd examples/kubeflow
docker build -t {MY-REGISTRY}/mnist-model -f Dockerfile.model .
docker push {MY-REGISTRY}/mnist-model
```

--------------------------------

### Configure Go GOPATH in VSCode

Source: https://github.com/datashim-io/datashim/blob/master/docs/GolangVSCodeGit.md

Add these settings to your VSCode JSON configuration to inform the Go extension about your Go installation directory and your custom development workspace.

```json
"go.toolsGopath": "$HOME/go",
"go.gopath": "$HOME/goprojects"
```

--------------------------------

### Create S3 Dataset Example

Source: https://github.com/datashim-io/datashim/blob/master/docs/Ceph-Caching.md

Defines an S3 dataset using the Dataset CRD. Replace placeholder values with your actual AWS credentials and S3 endpoint.

```yaml
apiVersion: datashim.io/v1alpha1
kind: Dataset
metadata:
  name: example-dataset
spec:
  local:
    type: "COS"
    accessKeyID: "{AWS_ACCESS_KEY_ID}"
    secretAccessKey: "{AWS_SECRET_ACCESS_KEY}"
    endpoint: "{S3_SERVICE_URL}"
    bucket: "{BUCKET_NAME}"
    region: "" #it can be empty
```

--------------------------------

### Configure Nooba for Datashim Hive

Source: https://github.com/datashim-io/datashim/blob/master/examples/hive/sampleapp/README.md

Set environment variables for Nooba as your Object Storage provider. Ensure S3_ENDPOINT is unset and NOOBAA_HOME is set to your Nooba installation path.

```bash
$ unset S3_ENDPOINT
$ export NOOBAA_HOME="path/to/noobaa/installation"
```

--------------------------------

### Create Datashim Demo Namespace

Source: https://github.com/datashim-io/datashim/blob/master/examples/inference-server/a-b-testing/README.md

Creates a Kubernetes namespace named 'datashim-demo' and sets it as the current context.

```command-line
kubectl create namespace datashim-demo
kubectl config set-context --current --namespace=datashim-demo
```

--------------------------------

### Clone Repository and Deploy

Source: https://github.com/datashim-io/datashim/wiki/Installation

Clone the dataset lifecycle framework repository and execute the deployment command.

```bash
git clone https://github.com/IBM/dataset-lifecycle-framework.git
cd dataset-lifecycle-framework
make deployment
```

--------------------------------

### Execute Datashim Hive Sample Application

Source: https://github.com/datashim-io/datashim/blob/master/examples/hive/sampleapp/README.md

Run the sample application script after configuring your environment and Docker registry. This will create a dataset, mount it, and display a CSV file.

```bash
$ ./run_sample.sh
```

--------------------------------

### Dataset Custom Resource Definition Example

Source: https://github.com/datashim-io/datashim/blob/master/src/dataset-operator/README.md

An example of a Dataset Custom Resource Definition (CRD) object. This defines a dataset with its type and configuration.

```yaml
apiVersion: datashim.io/v1alpha1
kind: Dataset
metadata:
  name: example-dataset
spec:
  # Add fields here
  type: COS
  conf:
    onefield: valueaa
    anotherfield: bb
```

--------------------------------

### Deploy Dataset Lifecycle Framework (DLF)

Source: https://github.com/datashim-io/datashim/wiki/Ceph-Caching

Builds and deploys the DLF using its Makefile. This command should be run from the root of the DLF directory.

```bash
make deployment
```

--------------------------------

### Create Dataset

Source: https://github.com/datashim-io/datashim/blob/master/docs/Ceph-Caching.md

Create a dataset using the provided YAML file after the plugin is deployed.

```bash
kubectl create -f my-dataset.yaml
```

--------------------------------

### Get CSI S3 Source Code

Source: https://github.com/datashim-io/datashim/blob/master/src/csi-s3/README.md

Download the CSI S3 source code using go get. This is a prerequisite for building the driver from source.

```bash
go get -u github.com/ctrox/csi-s3
```

--------------------------------

### Create Dataset Description for NooBaa

Source: https://github.com/datashim-io/datashim/wiki/Noobaa

Run this helper script to generate a Dataset yaml file for NooBaa. The script uses a template and populates it with connection details for NooBaa, creating a file named 'dataset-noobaa.yaml'.

```bash
./examples/noobaa/create_dataset_desc.sh
```

--------------------------------

### Inspect Mounted Dataset in Pod

Source: https://github.com/datashim-io/datashim/wiki/Archive-based-Datasets

Execute into a running pod and inspect the contents of the mounted dataset at '/mnt/datasets/example-dataset/'.

```bash
$ kubectl exec -it nginx /bin/bash
root@nginx:/# ls /mnt/datasets/example-dataset/
noaa-weather-data-jfk-airport
root@nginx:/# ls /mnt/datasets/example-dataset/noaa-weather-data-jfk-airport/
LICENSE.txt  README.txt  clean_data.py  jfk_weather.csv  jfk_weather_cleaned.csv
```

--------------------------------

### Tag and Load Scheduler Images

Source: https://github.com/datashim-io/datashim/blob/master/examples/scheduling/README.md

Tag Docker images for the scheduler and controller, then load them into Minikube.

```bash
$ docker tag localhost:5000/scheduler-plugins/kube-scheduler:latest datashim-test/kube-scheduler:cacheaware
$ docker tag localhost:5000/scheduler-plugins/controller:latest datashim-test/controller:cacheaware
$ minikube image load datashim-test/kube-scheduler:cacheaware
$ minikube image load datashim-test/controller:cacheaware
```

--------------------------------

### Get NodeID using CSC

Source: https://github.com/datashim-io/datashim/blob/master/src/csi-driver-nfs/README.md

Retrieve the Node ID of the CSI driver using the csc tool.

```shell
csc node get-id --endpoint tcp://127.0.0.1:10000
```

--------------------------------

### Configure and Install Ceph Operator

Source: https://github.com/datashim-io/datashim/wiki/Ceph-Caching

Applies the operator configuration for Rook Ceph. Ensure the `YOUR_REGISTRY` placeholder is replaced.

```bash
kubectl create -f operator.yaml
```

--------------------------------

### Create Test Pod

Source: https://github.com/datashim-io/datashim/blob/master/src/csi-s3/README.md

Deploy a test pod that mounts the PVC created earlier. This verifies that the CSI driver can successfully provision and attach the volume.

```bash
kubectl create -f poc.yaml
```

--------------------------------

### Get Plugin Info using CSC

Source: https://github.com/datashim-io/datashim/blob/master/src/csi-driver-nfs/README.md

Retrieve information about the NFS CSI plugin using the csc tool.

```shell
csc identity plugin-info --endpoint tcp://127.0.0.1:10000
```

--------------------------------

### Build NFS Plugin

Source: https://github.com/datashim-io/datashim/blob/master/src/csi-driver-nfs/README.md

Build the NFS plugin using the make command.

```shell
make nfs
```

--------------------------------

### Deploy MinIO Instance for Archive Datasets

Source: https://github.com/datashim-io/datashim/blob/master/docs/user-guide.md

Command to apply a MinIO instance configuration, which automatically creates a 'minio-conf' secret required for archive Datasets.

```bash
kubectl apply -n dlf -f https://github.com/datashim-io/datashim/raw/master/examples/minio/minio.yaml
```

--------------------------------

### Add Datashim Helm Repository

Source: https://github.com/datashim-io/datashim/blob/master/docs/index.md

Adds the Datashim Helm repository to your local Helm configuration. This is a prerequisite for installing Datashim using Helm.

```bash
helm repo add datashim https://datashim-io.github.io/datashim/
helm repo update
```

--------------------------------

### Create Pod Using NooBaa Dataset

Source: https://github.com/datashim-io/datashim/wiki/Noobaa

This Kubernetes Pod definition mounts the 'example-dataset' created from NooBaa into the Nginx container at '/mnt/datasets/example-dataset'. Changes made within this directory in the pod will be reflected in the NooBaa bucket.

```yaml
apiVersion: v1
kind: Pod
metadata:
  name: nginx
  labels:
    dataset.0.id: "example-dataset"
    dataset.0.useas: "mount"
spec:
  containers:
    - name: nginx
      image: nginx

```

--------------------------------

### Check Provisioner Logs

Source: https://github.com/datashim-io/datashim/blob/master/src/csi-s3/README.md

Troubleshoot PVC creation issues by checking the logs of the CSI provisioner. This can provide insights into errors during volume provisioning.

```bash
kubectl logs -l app=csi-provisioner-s3 -c csi-s3
```

--------------------------------

### Apply TGI Pod and Service Manifest

Source: https://github.com/datashim-io/datashim/blob/master/examples/inference-server/README.md

Apply the Kubernetes manifest to create the TGI pod and its associated service.

```bash
kubectl apply -f tgi+service.yaml
```

--------------------------------

### Apply MinIO Deployment

Source: https://github.com/datashim-io/datashim/blob/master/examples/inference-server/a-b-testing/README.md

Applies the MinIO deployment configuration. This is an optional step for setting up a local object storage endpoint.

```command-line
kubectl create namespace minio
kubectl apply -f minio.yaml
kubectl wait pod --for=condition=Ready -n minio --timeout=-1s minio
```

--------------------------------

### Create a Dataset

Source: https://github.com/datashim-io/datashim/blob/master/examples/scheduling/README.md

Define and apply a Dataset resource using kubectl. Ensure you replace placeholders like '<secret-name>', '<endpoint-name>', and '<bucket-name>' with your actual values.

```yaml
---
apiVersion: datashim.io/v1alpha1
kind: Dataset
metadata:
  name: example-dataset
spec:
  local:
    type: "COS"
    secret-name: "<secret-name>"
    endpoint: "<endpoint-name>"
    bucket: "<bucket-name>"
```

--------------------------------

### Create Tensorboard Deployment using kubectl

Source: https://github.com/datashim-io/datashim/wiki/Kubeflow-Model-Storing-and-Serving-with-DLF

Command to create the Tensorboard deployment and associated resources in your Kubernetes namespace.

```bash
kubectl create -f tensorboard.yaml -n {my-namespace}
```

--------------------------------

### Create and Verify Dataset Internal and PVC

Source: https://github.com/datashim-io/datashim/blob/master/docs/Ceph-Caching.md

Commands to create an S3 dataset and then verify its creation by checking the 'datasetsinternal' and PVC resources.

```bash
kubectl create -f my-dataset.yaml
kubectl get datasetsinternal
kubectl get pvc
```

--------------------------------

### Create TGI Deployments

Source: https://github.com/datashim-io/datashim/blob/master/examples/inference-server/simple-testing/README.md

Applies the TGI deployment configuration to both the 'production' and 'staging' namespaces. This deploys the Text Generation Inference server for serving AI models.

```bash
kubectl apply -n production -f deployment.yaml
kubectl apply -n staging -f deployment.yaml
```

--------------------------------

### Verify Pod Scheduling

Source: https://github.com/datashim-io/datashim/blob/master/examples/scheduling/README.md

Check the 'NODE' column in the output of 'kubectl get pods -o wide' to confirm that the 'testpod' has been scheduled to the node with the gateway for the cached dataset.

```bash
$ kubectl get pods -o wide
```

--------------------------------

### Unset S3 Endpoint and Export Noobaa Home

Source: https://github.com/datashim-io/datashim/blob/master/examples/hive/k8s/README.md

Configure ObjectStorage integration by unsetting the S3 endpoint and exporting the Noobaa home directory. Use this when employing the Noobaa install method.

```bash
$ unset S3_ENDPOINT
$ export NOOBAA_HOME=path/to/Noobaa/directory
```

--------------------------------

### Create TGI Deployments

Source: https://github.com/datashim-io/datashim/blob/master/examples/inference-server/a-b-testing/README.md

Applies Kubernetes configurations to create stable and canary deployments for Text Generation Inference (TGI) models. GPU selection can be configured via nodeSelector.

```command-line
kubectl apply -f tgi-stable.yaml
kubectl apply -f tgi-canary.yaml
```

```yaml
nodeSelector:
  nvidia.com/gpu.product: Tesla-T4
```

--------------------------------

### Create NooBaa Dataset

Source: https://github.com/datashim-io/datashim/wiki/Noobaa

Use this kubectl command to create the dataset pointer to the NooBaa bucket using the previously generated YAML file.

```bash
kubectl create -f ./examples/noobaa/dataset-noobaa.yaml
```

--------------------------------

### Wait for Datashim Pods to be Ready

Source: https://github.com/datashim-io/datashim/blob/master/README.md

Waits until all Datashim pods in the 'dlf' namespace are in a ready state.

```bash
kubectl wait --for=condition=ready pods -l app.kubernetes.io/name=datashim -n dlf
```

--------------------------------

### Create TensorFlow Serving Deployment

Source: https://github.com/datashim-io/datashim/wiki/Kubeflow-Model-Storing-and-Serving-with-DLF

Command to create the TensorFlow Serving deployment and service.

```bash
kubectl create -f tensorflow-serving -n {my-namespace}
```

--------------------------------

### Create KFServing InferenceService

Source: https://github.com/datashim-io/datashim/wiki/Kubeflow-Model-Storing-and-Serving-with-DLF

Command to create the KFServing InferenceService YAML definition.

```bash
kubectl create -f kfserving-inference.yaml -n {my-namespace}
```

--------------------------------

### Build CSI S3 Executable

Source: https://github.com/datashim-io/datashim/blob/master/src/csi-s3/README.md

Build the CSI S3 executable using the make command. This command compiles the Go source code into a runnable binary.

```bash
make build
```

--------------------------------

### Download and Deploy Models to MinIO

Source: https://github.com/datashim-io/datashim/blob/master/examples/inference-server/simple-testing/README.md

Applies Kubernetes Jobs to download FLAN-T5 models into the MinIO instance for both production and staging namespaces. It then waits for the download jobs to complete.

```bash
kubectl apply -f download-flan-t5-small-to-minio-prod.yaml
kubectl wait -n production --for=condition=complete job/download-flan --timeout=-1s
kubectl apply -f download-flan-t5-base-to-minio-staging.yaml
kubectl wait -n staging --for=condition=complete job/download-flan --timeout=-1s
```

--------------------------------

### Create Production and Staging Namespaces

Source: https://github.com/datashim-io/datashim/blob/master/examples/inference-server/simple-testing/README.md

Creates the 'production' and 'staging' namespaces in Kubernetes. These namespaces will be used to isolate different environments for model development.

```bash
kubectl create namespace production
kubectl create namespace staging
```

--------------------------------

### Clone and Checkout Archive Branch

Source: https://github.com/datashim-io/datashim/wiki/Archive-based-Datasets

Clone the dataset-lifecycle-framework repository and checkout the 'archive' branch.

```bash
git clone https://github.com/IBM/dataset-lifecycle-framework.git
cd dataset-lifecycle-framework
git checkout archive
```

--------------------------------

### Compile Scheduler Plugin Image

Source: https://github.com/datashim-io/datashim/blob/master/examples/scheduling/README.md

Compiles the scheduler plugin code and builds a local Docker image. Ensure GOPATH is set correctly before running.

```bash
export GOPATH=$DATASHIM_HOME/examples/scheduling/src/sigs.k8s.io/scheduler-plugins
make local-image
```

--------------------------------

### Download FLAN-T5 Model to MinIO

Source: https://github.com/datashim-io/datashim/blob/master/examples/inference-server/README.md

Applies a Kubernetes Job to download the FLAN-T5-Base model from Hugging Face and copy it into the specified Datashim Dataset mount path in MinIO. Waits for the job to complete.

```command-line
kubectl apply -f download-flan-t5-base-to-minio.yaml
kubectl wait --for=condition=complete job/download-flan --timeout=-1s
```

--------------------------------

### Test S3 Mount in Pod

Source: https://github.com/datashim-io/datashim/blob/master/src/csi-s3/README.md

Execute commands inside the test pod to verify the S3 volume is mounted correctly. This includes checking the mount point and creating a test file.

```bash
kubectl exec -ti csi-s3-test-nginx bash
$mount | grep fuse
$touch /var/lib/www/html/hello_world
```

--------------------------------

### Monitor TGI Pod Logs

Source: https://github.com/datashim-io/datashim/blob/master/examples/inference-server/README.md

Stream the logs from the TGI pod to observe its startup process and identify any potential issues.

```bash
kubectl logs -f text-generation-inference
```

--------------------------------

### Check CSI S3 Driver Logs

Source: https://github.com/datashim-io/datashim/blob/master/src/csi-s3/README.md

Troubleshoot issues related to container creation or volume mounting by examining the logs of the CSI S3 driver. Ensure the 'MountPropagation' feature gate is enabled.

```bash
kubectl logs -l app=csi-s3 -c csi-s3
```

--------------------------------

### Download Models to MinIO

Source: https://github.com/datashim-io/datashim/blob/master/examples/inference-server/a-b-testing/README.md

Applies Kubernetes Jobs to download FLAN-T5 models to the MinIO instance and waits for their completion. This may take several minutes.

```command-line
kubectl apply -f download-flan-t5-small-to-minio.yaml
kubectl wait --for=condition=complete job/download-flan-small --timeout=-1s
kubectl apply -f download-flan-t5-base-to-minio.yaml
kubectl wait --for=condition=complete job/download-flan-base --timeout=-1s
```

--------------------------------

### Deploy Datashim on Kubernetes/Minikube/kind

Source: https://github.com/datashim-io/datashim/blob/master/README.md

Applies the Datashim manifest for standard Kubernetes environments.

```bash
kubectl apply -f https://raw.githubusercontent.com/datashim-io/datashim/master/release-tools/manifests/dlf.yaml
```

--------------------------------

### Build E2E Test Binary

Source: https://github.com/datashim-io/datashim/blob/master/src/csi-driver-nfs/README.md

Build the Kubernetes End-to-End test binary for the CSI NFS driver.

```shell
make build-tests
```

--------------------------------

### Create TGI Services

Source: https://github.com/datashim-io/datashim/blob/master/examples/inference-server/simple-testing/README.md

Applies the TGI service configuration to both the 'production' and 'staging' namespaces. This creates network endpoints for accessing the deployed TGI models.

```bash
kubectl apply -n production -f service.yaml
kubectl apply -n staging -f service.yaml
```

--------------------------------

### Set Up Upstream Remote and Rebase

Source: https://github.com/datashim-io/datashim/blob/master/docs/GitWorkflow.md

Configure the official Datashim repository as an upstream remote and ensure your local master branch is up-to-date.

```bash
$
$> cd $HOME/goprojects/src/datashim
$> git remote add upstream https://github.com/datashim-io/datashim
$> git remote set-url --push upstream no_push 
$> git fetch upstream
$> git checkout master
$> git rebase upstream/master
```

--------------------------------

### Deploy Datashim

Source: https://github.com/datashim-io/datashim/blob/master/examples/scheduling/README.md

Deploys Datashim to the Kubernetes cluster. This command should be run from the Datashim root directory.

```bash
cd $DATASHIM_HOME && make deployment
```

--------------------------------

### Describe Dataset Status

Source: https://github.com/datashim-io/datashim/blob/master/examples/scheduling/README.md

Examine the status of the 'example-dataset', paying attention to the 'Provision.Status' which should be 'OK'. The 'Caching.Status' may show 'Pending' initially.

```bash
$ kubectl describe dataset example-dataset
```

--------------------------------

### Wait for TGI Pod to be Ready

Source: https://github.com/datashim-io/datashim/blob/master/examples/inference-server/README.md

Monitor the Kubernetes pod until it reaches the 'Ready' condition. This command will block until the pod is ready.

```bash
kubectl wait pod --for=condition=Ready text-generation-inference --timeout=-1s
```

--------------------------------

### Run CSI S3 Driver Tests

Source: https://github.com/datashim-io/datashim/blob/master/src/csi-s3/README.md

Execute the end-to-end tests for the CSI S3 driver using the provided Makefile. This command initiates the testing process within a Docker container, which requires S3 storage and a mounter like s3fs.

```bash
make test
```

--------------------------------

### Mount Dataset in a Pod using Labels

Source: https://github.com/datashim-io/datashim/blob/master/README.md

Configures a pod to mount a defined Dataset by using labels. The Dataset will be available at /mnt/datasets/example-dataset.

```yaml
apiVersion: v1
kind: Pod
metadata:
  name: nginx
  labels:
    dataset.0.id: "example-dataset"
    dataset.0.useas: "mount"
spec:
  containers:
    - name: nginx
      image: nginx
```

--------------------------------

### Wait for TGI Deployments to be Ready

Source: https://github.com/datashim-io/datashim/blob/master/examples/inference-server/a-b-testing/README.md

Waits for the stable and canary Text Generation Inference (TGI) pods to reach a Ready state.

```command-line
kubectl wait pod -l run=text-generation-inference -l type=stable --for=condition=Ready --timeout=-1s
kubectl wait pod -l run=text-generation-inference -l type=canary --for=condition=Ready --timeout=-1s
```

--------------------------------

### View Scheduler Pod Logs

Source: https://github.com/datashim-io/datashim/blob/master/examples/scheduling/README.md

Inspect the logs of the scheduler plugin pod to verify the scheduling decisions, including dataset caching status and node assignments.

```bash
$ kubectl logs -n scheduler-plugins scheduler-plugins-scheduler-67b979c8db-ml6lm
```

--------------------------------

### Deploy Datashim on OpenShift

Source: https://github.com/datashim-io/datashim/blob/master/README.md

Applies the Datashim manifest for OpenShift environments.

```bash
kubectl apply -f https://raw.githubusercontent.com/datashim-io/datashim/master/release-tools/manifests/dlf-oc.yaml
```

--------------------------------

### Label Namespace for Datashim

Source: https://github.com/datashim-io/datashim/blob/master/examples/inference-server/a-b-testing/README.md

Labels the 'datashim-demo' namespace with 'monitor-pods-datasets=enabled' to allow Datashim to mount volumes in pods.

```command-line
kubectl label namespace datashim-demo monitor-pods-datasets=enabled
```

--------------------------------

### Create TFJob using kubectl

Source: https://github.com/datashim-io/datashim/wiki/Kubeflow-Model-Storing-and-Serving-with-DLF

Command to create the TFJob in your specified Kubernetes namespace.

```bash
kubectl create -f tfjob.yaml -n {my-namespace}
```

--------------------------------

### Create Persistent Volume Claim (PVC)

Source: https://github.com/datashim-io/datashim/blob/master/src/csi-s3/README.md

Create a PVC using the previously defined storage class. This requests a persistent volume from the CSI S3 driver.

```bash
kubectl create -f pvc.yaml
```

--------------------------------

### Launch TFJob with DLF Mounted Dataset

Source: https://github.com/datashim-io/datashim/wiki/Kubeflow-Model-Storing-and-Serving-with-DLF

Use this YAML to define a TFJob that automatically mounts a specified dataset via DLF. Replace 'your-dataset' with your actual dataset name.

```yaml
apiVersion: kubeflow.org/v1
kind: TFJob
metadata:
  name: my-train
spec:
  tfReplicaSpecs:
    Ps:
      replicas: 1
      template:
        metadata:
          labels:
           dataset.0.id: "your-dataset"
           dataset.0.useas: "mount"
          annotations:
            sidecar.istio.io/inject: "false"
        spec:
          serviceAccount: default-editor
          containers:
            - name: tensorflow
              command:
                - python
                - /opt/model.py
                - --tf-model-dir=/mnt/datasets/your-dataset/mnist
                - --tf-export-dir=/mnt/datasets/your-dataset/mnist/export
                - --tf-train-steps=200
                - --tf-batch-size=100
                - --tf-learning-rate=0.1
              image: yiannisgkoufas/mnist
              workingDir: /opt
              resources:
                limits:
                  ephemeral-storage: "10Gi"
                requests:
                  ephemeral-storage: "10Gi"
          restartPolicy: OnFailure
    Chief:
      replicas: 1
      template:
        metadata:
          labels:
            dataset.0.id: "your-dataset"
            dataset.0.useas: "mount"
          annotations:
            sidecar.istio.io/inject: "false"
        spec:
          serviceAccount: default-editor
          containers:
            - name: tensorflow
              resources:
                limits:
                  ephemeral-storage: "10Gi"
                requests:
                  ephemeral-storage: "10Gi"
              command:
                - python
                - /opt/model.py
                - --tf-model-dir=/mnt/datasets/your-dataset/mnist
                - --tf-export-dir=/mnt/datasets/your-dataset/mnist/export
                - --tf-train-steps=200
                - --tf-batch-size=100
                - --tf-learning-rate=0.1
              image: yiannisgkoufas/mnist
          restartPolicy: OnFailure
    Worker:
      replicas: 1
      template:
        metadata:
          labels:
            dataset.0.id: "your-dataset"
            dataset.0.useas: "mount"
          annotations:
            sidecar.istio.io/inject: "false"
        spec:
          serviceAccount: default-editor
          containers:
            - name: tensorflow
              command:
                - python
                - /opt/model.py
                - --tf-model-dir=/mnt/datasets/your-dataset/mnist
                - --tf-export-dir=/mnt/datasets/your-dataset/mnist/export
                - --tf-train-steps=200
                - --tf-batch-size=100
                - --tf-learning-rate=0.1
              image: yiannisgkoufas/mnist
              workingDir: /opt
          restartPolicy: OnFailure

```

--------------------------------

### Search for Intermediate Datashim Helm Releases

Source: https://github.com/datashim-io/datashim/blob/master/README.md

Searches the Datashim Helm repository for intermediate releases (e.g., alpha, beta) using the --devel flag.

```bash
helm search repo datashim --devel
```

--------------------------------

### Wait for TGI Deployments to be Ready

Source: https://github.com/datashim-io/datashim/blob/master/examples/inference-server/simple-testing/README.md

Waits for the Text Generation Inference (TGI) pods to become ready in both the 'production' and 'staging' namespaces. This ensures the TGI servers are operational before proceeding.

```bash
kubectl wait pod -n production -l run=text-generation-inference --for=condition=Ready --timeout=-1s
kubectl wait pod -n staging -l run=text-generation-inference --for=condition=Ready --timeout=-1s
```

--------------------------------

### Apply Datashim Dataset Configurations

Source: https://github.com/datashim-io/datashim/blob/master/examples/inference-server/simple-testing/README.md

Applies Kubernetes configurations for S3 secrets and Datasets in both production and staging environments. This step makes the data accessible via Datashim.

```bash
kubectl apply -f s3-secret-prod.yaml
kubectl apply -f dataset-prod.yaml
kubectl apply -f s3-secret-staging.yaml
kubectl apply -f dataset-staging.yaml
```

--------------------------------

### Verify DLF Pods

Source: https://github.com/datashim-io/datashim/wiki/Ceph-Caching

Lists pods in the default namespace to confirm the successful deployment of DLF components.

```bash
csi-attacher-nfsplugin-0            2/2     Running   0          7s
csi-attacher-s3-0                   1/1     Running   0          8s
csi-nodeplugin-nfsplugin-nqgtl      2/2     Running   0          7s
csi-provisioner-s3-0                2/2     Running   0          8s
csi-s3-k9b5j                        2/2     Running   0          8s
dataset-operator-7b8f65f7d4-hg8n5   1/1     Running   0          6s
```

--------------------------------

### Kubernetes Feature Gates and Runtime Config

Source: https://github.com/datashim-io/datashim/blob/master/src/csi-driver-nfs/README.md

Enable necessary feature gates and runtime configurations for the CSI driver. Mount propagation requires privileged containers.

```shell
FEATURE_GATES=CSIPersistentVolume=true,MountPropagation=true
RUNTIME_CONFIG="storage.k8s.io/v1alpha1=true"
```