Azure¶

You can deploy Pachyderm in a new or existing Microsoft® Azure® Kubernetes Service environment and use Azure's resource to run your Pachyderm workloads. To deploy Pachyderm to AKS, you need to:

1. Install Prerequisites
2. Deploy Kubernetes
3. Deploy Pachyderm

Install Prerequisites¶

Before you can deploy Pachyderm on Azure, you need to configure a few prerequisites on your client machine. If not explicitly specified, use the latest available version of the components listed below. Install the following prerequisites:

• Azure CLI 2.0.1 or later
• jq
• kubectl
• pachctl

Install pachctl¶

pachctl is a primary command-line utility for interacting with Pachyderm clusters. You can run the tool on Linux®, macOS®, and Microsoft® Windows® 10 or later operating systems and install it by using your favorite command line package manager. This section describes how you can install pachctl by using brew and curl.

If you are installing pachctl on Windows, you need to first install Windows Subsystem (WSL) for Linux.

To install pachctl, complete the following steps:

• To install on macOS by using brew, run the following command:

  $ brew tap pachyderm/tap && brew install pachyderm/tap/pachctl@1.9
  

  * To install on Linux 64-bit or Windows 10 or later, run the following command:

  $ curl -o /tmp/pachctl.deb -L https://github.com/pachyderm/pachyderm/releases/download/v1.9.9/pachctl_1.9.9_amd64.deb &&  sudo dpkg -i /tmp/pachctl.deb
  

1. Verify your installation by running pachctl version:

   $ pachctl version --client-only
   COMPONENT           VERSION
   pachctl             1.9.0
   

Deploy Kubernetes¶

You can deploy Kubernetes on Azure by following the official Azure Container Service documentation or by following the steps in this section. When you deploy Kubernetes on Azure, you need to specify the following parameters:

To deploy Kubernetes on Azure, complete the following steps:

1. Log in to Azure:

   $ az login
   Note, we have launched a browser for you to login. For old experience with
   device code, use "az login --use-device-code"
   

   If you have not already logged in this command opens a browser window. Log in with your Azure credentials. After you log in, the following message appears in the command prompt:

   You have logged in. Now let us find all the subscriptions to which you have access...
   [
     {
       "cloudName": "AzureCloud",
       "id": "your_id",
       "isDefault": true,
       "name": "Microsoft Azure Sponsorship",
       "state": "Enabled",
       "tenantId": "your_tenant_id",
       "user": {
         "name": "your_contact_id",
         "type": "user"
       }
     }
   ]
   

2. Create an Azure resource group.

   $ az group create --name=${RESOURCE_GROUP} --location=${LOCATION}
   

   Example:

   $ az group create --name="test-group" --location=centralus
   {
     "id": "/subscriptions/6c9f2e1e-0eba-4421-b4cc-172f959ee110/resourceGroups/pach-resource-group",
     "location": "centralus",
     "managedBy": null,
     "name": "pach-resource-group",
     "properties": {
       "provisioningState": "Succeeded"
     },
     "tags": null,
     "type": null
   }
   

3. Create an AKS cluster:

   $ az aks create --resource-group ${RESOURCE_GROUP} --name ${CLUSTER_NAME} --generate-ssh-keys --node-vm-size ${NODE_SIZE}
   

   Example:

   $ az aks create --resource-group test-group --name test-cluster --generate-ssh-keys --node-vm-size Standard_DS4_v2
   {
     "aadProfile": null,
     "addonProfiles": null,
     "agentPoolProfiles": [
       {
         "availabilityZones": null,
         "count": 3,
         "enableAutoScaling": null,
         "maxCount": null,
         "maxPods": 110,
         "minCount": null,
         "name": "nodepool1",
         "orchestratorVersion": "1.12.8",
         "osDiskSizeGb": 100,
         "osType": "Linux",
         "provisioningState": "Succeeded",
         "type": "AvailabilitySet",
         "vmSize": "Standard_DS4_v2",
         "vnetSubnetId": null
       }
     ],
   ...
   

4. Confirm the version of the Kubernetes server:

   $ kubectl version
   Client Version: version.Info{Major:"1", Minor:"13", GitVersion:"v1.13.4", GitCommit:"c27b913fddd1a6c480c229191a087698aa92f0b1", GitTreeState:"clean", BuildDate:"2019-03-01T23:36:43Z", GoVersion:"go1.12", Compiler:"gc", Platform:"darwin/amd64"}
   Server Version: version.Info{Major:"1", Minor:"13", GitVersion:"v1.13.4", GitCommit:"c27b913fddd1a6c480c229191a087698aa92f0b1", GitTreeState:"clean", BuildDate:"2019-02-28T13:30:26Z", GoVersion:"go1.11.5", Compiler:"gc", Platform:"linux/amd64"}
   

See also:

• Azure Virtual Machine sizes

Add storage resources¶

Pachyderm requires you to deploy an object store and a persistent volume in your cloud environment to function correctly. For best results, you need to use faster disk drives, such as Premium SSD Managed Disks that are available with the Azure Premium Storage offering.

You need to specify the following parameters when you create storage resources:

To create these resources, follow these steps:

1. Clone the Pachyderm GitHub repo.
2. Change the directory to the root directory of the pachyderm repository.

3. Create an Azure storage account:

   $ az storage account create \
     --resource-group="${RESOURCE_GROUP}" \
     --location="${LOCATION}" \
     --sku=Premium_LRS \
     --name="${STORAGE_ACCOUNT}" \
     --kind=BlockBlobStorage
   

   System response:

   {
     "accessTier": null,
     "creationTime": "2019-06-20T16:05:55.616832+00:00",
     "customDomain": null,
     "enableAzureFilesAadIntegration": null,
     "enableHttpsTrafficOnly": false,
     "encryption": {
       "keySource": "Microsoft.Storage",
       "keyVaultProperties": null,
       "services": {
         "blob": {
           "enabled": true,
     ...
   

   Make sure that you set Stock Keeping Unit (SKU) to Premium_LRS and the kind parameter is set to BlockBlobStorage. This configuration results in a storage that uses SSDs rather than standard Hard Disk Drives (HDD). If you set this parameter to an HDD-based storage option, your Pachyderm cluster will be too slow and might malfunction.

4. Verify that your storage account has been successfully created:

   $ az storage account list
   

5. Build a Microsoft tool for creating Azure VMs from an image:

   $ STORAGE_KEY="$(az storage account keys list \
                 --account-name="${STORAGE_ACCOUNT}" \
                 --resource-group="${RESOURCE_GROUP}" \
                 --output=json \
                 | jq '.[0].value' -r
              )"
   

6. Find the generated key in the Storage accounts > Access keys section in the Azure Portal or by running the following command:

   $ az storage account keys list --account-name=${STORAGE_ACCOUNT}
   [
     {
       "keyName": "key1",
       "permissions": "Full",
       "value": ""
     }
   ]
   

See Also

• Azure Storage

Deploy Pachyderm¶

After you complete all the sections above, you can deploy Pachyderm on Azure. If you have previously tried to run Pachyderm locally, make sure that you are using the right Kubernetes context. Otherwise, you might accidentally deploy your cluster on Minikube.

1. Verify cluster context:

   $ kubectl config current-context
   

   This command should return the name of your Kubernetes cluster that runs on Azure.

   • If you have a different contents displayed, configure kubectl to use your Azure configuration:

   $ az aks get-credentials --resource-group ${RESOURCE_GROUP} --name ${CLUSTER_NAME}
   Merged "${CLUSTER_NAME}" as current context in /Users/test-user/.kube/config
   

2. Run the following command:

   $ pachctl deploy microsoft ${CONTAINER_NAME} ${STORAGE_ACCOUNT} ${STORAGE_KEY} ${STORAGE_SIZE} --dynamic-etcd-nodes 1
   

   Example:

   $ pachctl deploy microsoft test-container teststorage <key> 10 --dynamic-etcd-nodes 1
   serviceaccount/pachyderm configured
   clusterrole.rbac.authorization.k8s.io/pachyderm configured
   clusterrolebinding.rbac.authorization.k8s.io/pachyderm configured
   service/etcd-headless created
   statefulset.apps/etcd created
   service/etcd configured
   service/pachd configured
   deployment.apps/pachd configured
   service/dash configured
   deployment.apps/dash configured
   secret/pachyderm-storage-secret configured
   
   Pachyderm is launching. Check its status with "kubectl get all"
   Once launched, access the dashboard by running "pachctl port-forward"
   

   Because Pachyderm pulls containers from DockerHub, it might take some time before the pachd pods start. You can check the status of the deployment by periodically running kubectl get all.

3. When pachyderm is up and running, get the information about the pods:

   $ kubectl get pods
   NAME                      READY     STATUS    RESTARTS   AGE
   dash-482120938-vdlg9      2/2       Running   0          54m
   etcd-0                    1/1       Running   0          54m
   pachd-1971105989-mjn61    1/1       Running   0          54m
   

   Note: Sometimes Kubernetes tries to start pachd nodes before the etcd nodes are ready which might result in the pachd nodes restarting. You can safely ignore those restarts.

4. To connect to the cluster from your local machine, such as your laptop, set up port forwarding to enable pachctl and cluster communication:

   $ pachctl port-forward
   

5. Verify that the cluster is up and running:

   $ pachctl version
   COMPONENT           VERSION
   pachctl             1.9.0
   pachd               1.9.0