AWS - Deploy the MCM and a Cluster

Introduction

VAST on Cloud (VoC) clusters are deployed using the Multi-Cluster Manager (MCM) cloud-based service. This service is provisioned via a VAST Data-provided AWS CloudFormation template. Each VoC cluster is automatically configured with a management access IP and predefined VIP pools for both replication and protocol access. This streamlined deployment process enables you to begin running workloads with minimal setup time.

Prerequisites

You will need to acquire a License Token to use the MCM and create clusters. Please contact the VAST on Cloud Sales Team to request one.

• NOTE: Tokens are single-use and only last 14 days from the date of issue. If you delete the MCM, you will need a new one.

Installation:

High-Level Steps:

1. Review and verify Prerequisites.

2. Request Token.

3. Deploy the Validation CloudFormation Stack (see document above).

   1. Resolve any issues.

4. Add a subscription to the VAST Data Platform in the Marketplace.

5. Launch the VAST Data CloudFormation Stack from the subscription page.

   1. Go through the launch steps.

   2. Wait for completion.

   3. Click on the MCM URL.

6. Log in to MCM.

   1. Apply Token.

7. Launch Cluster.

   1. Login

8. Configure and use a cluster.

NOTE:

• The installation starts at step 5.  Instructions for running the Validation Stack are in the attached document.

MCM Deployment from the AWS Marketplace

1. Launch the VAST Data CloudFormation Stack from the subscription page.

After you have subscribed to the VAST Data Platform, you will see it in your list of subscriptions. This account only has the one subscription - do not subscribe to the 4.7 version..

2. Choose the VAST Version, then → “Continue to Launch”.

3. Launch the CloudFormation Stack:

4. Create the Stack

Nothing to choose here, everything is pre-selected, and you will see the URL for the CloudFormation Stack prepopulated [Next]:

5. Specify Stack Details:

• Parameters -

  • Stack name: The name that will be used for the Multi-Cluster Manager Instance. You will live with this for a while, so choose wisely.

    • Required -

      • EnableCallHome:  <boolean> Can be enabled later.  

      • KeyName: <dropdown> Select the Key Pair you will use for SSH access.

      • SecurityGroupIds: <dropdown> Select an SG(s) configured for the subnets/VPC you are using.

      • DBSubnetsGroup: <dropdown> Select 2 subnets in different availability zones.

      • BucketName: Existing S3 bucket for MCM CloudWatch logs.

    • Optional -

      • AdditionalTags:  For additional tags your organization may require.

      • ExtraParams: <Reserved for VAST Engineering/Support>

      • DbSecurityGroupId: <Use the same SG chosen above>

    • Provided -

      • ImageId: Do not change this.

    • [Next]

6. Configure Stack Options

Acknowledge that IAM resources may get created [Next]

7. Review and Create

[Submit] starts the deployment.

8. CloudFormation Stacks - [Events]

Once you submit the stack, the console will switch to a different view that will scroll events as the Stack is provisioned. This is where you will see errors related to IAM permissions.

Watch the events until it reaches “CREATE_COMPLETE”

9. CloudFormation Stack [Outputs]

After the provisioning is complete, switch to the “Outputs” tab, where you will see a link for the MCM.

Clicking on the link will launch the MCM Management Console.

If you don’t have issues with insufficient privileges, you will get this far, and if you have a route to the VPC/subnet, you will be able to launch the MCM and log in.

At this point, you have successfully deployed the Multi Cluster Manager and will use your one-time token to register the new instance and start creating clusters.

10. Accessing the MCM: Logging In and Applying Token

• First, you get the usual unsigned certificate screen - click on “proceed to <stack name>:

• EULA:

• Login - standard default VAST Login <admin> <123456>:

• Once you are logged in, you will see the “Provide a registration token link at the top - click on it and type in the token you were given:

• Now you are fully up and running and ready to start up a cluster or 2:

Creating a Cluster

11. Cluster Creation from MCM

• Once you choose one of the options above, you will be placed in the “Create a Cluster' dialogue:

Notes:

• Naming follows the standard AWS naming conventions for resources.

• The region will be pre-chosen based on the VPC you are using.

• On Demand is recommended - Spot is cheaper but could be preempted.

• One more step: Clusters are installed by launching another CloudFormation Stack. It is triggered when you “Click to finish setup in AWS”. The next section will go over configuring and launching the stack.

12. Cluster Stack Configuration

When you “Set” the cluster parameters, you get switched to this AWS dialogue:

All of the defaults should be sufficient - the drop-downs will be pre-populated lists based on the MCM stack inputs.

ℹ️ Info

Be careful with the “create new Security Group” option - if you know your user has the permissions, you can, but if you aren’t sure, it's better to use the one already chosen/created.

• This should look familiar by now - time to go get some coffee.

• When it is done, there will be a new stack added - the “wrapper”. Select that one and the “Output” tab, and you will see “ClusterMgmt” - the value is the link to the VMS for that cluster.  Save it for later if you want to access the cluster directly without the MCM.

• And there it is - the standard new cluster login with the unsigned cert.

• Go back to the MCM - you will now see a new view with the cluster added in a pane on the right. This is how you will manage the cluster(s) you create:

• You can:

  • Open the VMS GUI

  • Remove the Cluster

  • Suspend the cluster

13. VMS Login/Running

We will skip the next few screens; you should be very familiar with them by now, and they are self-explanatory.

If everything has gone smoothly up to this point, you will have a running cluster and a familiar VMS GUI interface:

What now?

Now that the cluster is up and running, you have a fully functional 5.3 VAST Data cluster in a VPC. You can start creating cluster resources like views, DNS, VIPs, etc.

Guidance:

Deleting a Cluster and associated Resources

• This should be done through the MCM using “Remove Cluster.”

• You could do it by deleting the CloudFormation Stack, but it is cleaner and supported to use the MCM.

Deleting an MCM and associated Resources

• Go back to the CloudFormation Stack portal and “delete” the stack.

• If you left clusters behind, you can delete them from the same place by deleting their stacks.

Default cluster resources

• 2 VIP Pools with IPs pulled from the subnet the cluster lives in.

  • replicationPool

  • protocolsPool

• Policies -

  • default - NFS

  • s3_default_policy

Accessing and managing the cluster(s)

There are several ways to access a cluster once it is running.

Through the MCM

Once a cluster has been deployed, you can do some basic management from the interface.

Directly from the URL

This can be retrieved from the “Outputs” tab of the CloudFormation Stack that created the cluster.

You don’t need to be authenticated to AWS to access the MCM or cluster directly, so it is probably a good idea to save these URLs somewhere.

SSH login to the VM and use VCLI

You will use the .pem or .ppk key you created and used with the CloudFormation Stack that created the MCM/Cluster. You can add other keys later, but you will need to use the one you chose initially.

NOTE: You will need to have a route to the private subnet.

• Find the private IP of the Instance (or DNS name if you can resolve AWS private DNS:)

• Locate your .pem file

• SSH to the instance -

• Recommendation: either save the path in a variable for reuse, or better yet, add the entry to your .ssh config file.

.ssh/config:

Host source
    HostName 10.105.38.6
    User ubuntu
    IdentityFile C:\Users\karlv\.ssh\other_keys\aws.karlv-poc.pem
    PasswordAuthentication no

Host clustera
    HostName 10.105.28.77
    User centos
    IdentityFile C:\Users\karlv\.ssh\other_keys\aws.karlv-poc.pem
    PasswordAuthentication no

This lets you use the Host alias - “source” or “clustera” in this case directly -

KV C:\> ssh clustera
Last login: Fri Jul 18 22:34:10 2025 from ip-10-241-248-45.ec2.internal
   ,     #_
   ~\_  ####_        Amazon Linux 2
  ~~  \_#####\
  ~~     \###|       AL2 End of Life is 2026-06-30.
  ~~       \#/ ___
   ~~       V~' '->
    ~~~         /    A newer version of Amazon Linux is available!
      ~~._.   _/
         _/ _/       Amazon Linux 2023, GA and supported until 2028-03-15.
       _/m/'           https://aws.amazon.com/linux/amazon-linux-2023/

60 package(s) needed for security, out of 77 available
Run "sudo yum update" to apply all updates.
[centos@ip-10-105-28-243 ~]$

Using a variable -

ssh -i $AWSPem centos@10.105.28.77

If you aren’t familiar with the .ssh/config file, it is worth some investigation.

Use Terraform to configure clusters.

Complete Example

Resources:

• VAST on Cloud Overview

• VAST on Cloud for AWS

• VAST on Cloud for GCP