ACS-Engine with DCOS and Portworx

Overview

The Azure Container Service Engine (acs-engine) generates ARM (Azure Resource Manager) templates for Docker enabled clusters on Microsoft Azure with your choice of DC/OS, Kubernetes, Swarm Mode, or Swarm orchestrators. The input to the tool is a cluster definition. The cluster definition is very similar to (in many cases the same as) the ARM template syntax used to deploy a Microsoft Azure Container Service cluster.

The cluster definition file enables the following customizations to your Docker enabled cluster:

• choice of DC/OS, Kubernetes, Swarm Mode, or Swarm orchestrators
• multiple agent pools where each agent pool can specify:
• standard or premium VM Sizes,
• node count,
• Virtual Machine ScaleSets or Availability Sets,
• Storage Account Disks or Managed Disks (under private preview),
• Docker cluster sizes of 1200

The instructions below are presented only as a template for how to deploy Portworx on ACS-Engine using DCOS.

Install acs-engine and azure CLI

Install and build the acs-engine binary

From a Linux host:

• clone the repo:

git clone https://github.com/Azure/acs-engine.git

• cd into acs-engine

cd acs-engine

• install acs-engine

./scripts/devenv.sh && make prereqs && make build

• install Azure CLI:

curl -L https://aka.ms/InstallAzureCli | bash

Login to Azure and Set Subscription

az login

az account set --subscription "Your-Azure-Subscription-UUID"

Create Azure Resource Group and Location

Pick a name for the Azure Resource Group and choose a LOCATION value among the following: centralus,eastasia,southeastasia,eastus,eastus2,westus,westus2,northcentralus
southcentralus,westcentralus,northeurope,westeurope,japaneast,japanwest
brazilsouth,australiasoutheast,australiaeast,westindia,southindia,centralindia
canadacentral,canadaeast,uksouth,ukwest,koreacentral,koreasouth

Then, run the following command:

az group create --name "$RGNAME" --location "$LOCATION"

Select and customize the deployment configuration

The example deployment here uses DCOS with pre-attached disks and VM scale sets. The sample json file can be found in the acs-engine repository under examples/disks-managed/dcos-preAttachedDisks-vmss.json

The most important consideration for Portworx is to ensure that the target nodes have at least one “local” attached disk that can be used to contribute storage to the global storage pool.

For the masterProfile, specify an appropriate value for dnsPrefix which will be used for fully qualified domain name (FQDN).
Use the default vmSize or select an appropriate value for the machine type and size.
Specify the number and size of disks that will be attached to each DCOS private agent as per the template default:

[...]
"diskSizesGB": [128, 128, 128, 128]
[...]

Specify the appropriate admin username as adminUsername and public key data as keyData

Generate the Azure Resource Management (ARM) templates

./acs-engine generate examples/disks-managed/my-dcos-preAttachedDisks-vmss.json

The template will get generated under the acs-engine _output/$NAME where $NAME correspods to the name used for the dnsPrefix. acs-engine will generate the appropriate files for apimodel.json, azuredeploy.json, and azuredeploy.parameters.json

Deploy the generated ARM template

az group deployment create \
    --name "$NAME" \
    --resource-group "$RGNAME" \
    --template-file "./_output/$NAME/azuredeploy.json" \
    --parameters "./_output/$NAME/azuredeploy.parameters.json"

where $RGNAME corresponds to the resource group name created above, and $NAME corresonds to the above value used for dnsPrefix

Create an ssh tunnel to the DCOS GUI

Currently, ARM templates deployed through ACS-engine require that an ‘ssh’ tunnel be established to connect to the instance specified as the FQDN. As such for DCOS, the easiest way to manage the cluster is to establish an ssh tunnel that can be used to access the DCOS GUI. Full instructions for recommended approaches can be found here )

A simple example for access might look like this:

sudo ssh -i your-private.key  -fNL 80:localhost:80 -p 2200  \
     adminUsername@dnsPrefix.LOCATION.cloudapp.azure.com

where your-private.key is the private key that corresponds to keyData in the template
adminUsername and dnsPrefix refer to the corresponding values in the template
LOCATION corresponds to the location used for the resource group created.

Once the DCOS GUI is accessible, then install the DCOS CLI on your local machine

Install Portworx

Use the Run Portworx with Mesosphere DC/OS for installing the Portworx Frameworks on DCOS.

Once Portworx is installed, then the Portworx Stateful Service Frameworks can be easily deployed as per the reference documentation.