Click the Deploy to Azure button below to deploy the cloud resources on Azure®. Doing so opens the Azure Portal in your web browser.
Cluster Platform: Windows Server 2025 (MATLAB R2026a), Windows Server 2022 (MATLAB R2025b and earlier)
MATLAB® Release: R2025b
Clicking the Deploy to Azure button opens the "Custom deployment" page in your browser. You can configure the parameters on this page. It is easier to complete the steps if you position these instructions and the Azure Portal window side by side. Create a new resource group by clicking Create New. Alternatively, you can select an existing resource group, but doing so can cause conflicts if resources are already deployed in it.
- Specify and check the default settings for these resource parameters.
| Parameter Label | Description |
|---|---|
| Cluster Name | Name to use for this cluster. This name is shown in MATLAB as the cluster profile name. |
| Num Worker Nodes | Number of Azure virtual machines to start for the workers to run on. |
| Min Worker Nodes | Minimum number of running Azure virtual machines. |
| Max Worker Nodes | Maximum number of running Azure virtual machines. |
| Num Workers Per Node | Number of MATLAB worker processes to start on each virtual machine (VM). Specify 1 worker for every 2 vCPUs so that each physical core has 1 worker. For example, a Standard_D64s_v3 VM has 64 vCPUs, so it can support 32 MATLAB workers. See https://learn.microsoft.com/azure/virtual-machines/sizes for details on vCPUs for each VM size. |
| Head Node Vm Size | Azure virtual machine (VM) size to use for the head node, which runs the job manager. No workers are started on this node, so this node can be a smaller VM type than the worker nodes. By default, the heap memory for the job manager is set between 1024 MiB and a maximum of half of the VM memory, depending on the total number of MATLAB workers. See https://learn.microsoft.com/azure/virtual-machines/sizes for a list of VMs. |
| Worker Vm Size | Azure virtual machine (VM) size to use for the workers. By default, the heap memory for all worker process is set between 1024 MiB and a maximum of a quarter of the VM memory, depending on the number of MATLAB workers on the VM. See https://learn.microsoft.com/azure/virtual-machines/sizes for a list of VMs. |
| Use Spot Instances For Worker Nodes | Option indicating whether to enable Azure Spot Virtual Machines for worker nodes. For more information, refer to the FAQ section in the deployment README. |
| Database Volume Size | Size in GB of the volume to store the database files. All job and task information, including input and output data, is stored on this volume and so it must have enough capacity to store the expected amount of data. If set to 0, no volume is created and the root volume of the VM is used for the database. |
| Virtual Network Resource ID | (Optional) Resource ID of an existing virtual network to deploy your cluster into. You can find this under the Properties of your virtual network. If left empty, a new virtual network with a default subnet is created. |
| Subnet Name | (Optional) Name of an existing subnet within your chosen virtual network to deploy your cluster into. Required if a Virtual Network Resource ID is specified. |
| New Vnet Address Space | Address space for the new virtual network, specified in CIDR notation (e.g., 10.0.0.0/16). Only used if Virtual Network Resource ID is left empty. |
| New Subnet Address Space | Address space for the new subnet, specified in CIDR notation (e.g., 10.0.0.0/24). Only used if Virtual Network Resource ID is left empty. |
| Create Public IP Address | Choose whether to attach Public IPv4 addresses to the head-node and worker nodes. For details about using a private network configuration, see Configure Private Network. |
| Client IP Address List | Comma-separated list of IPv4 address ranges that can be used to access the cluster from MATLAB. This must be a valid IP CIDR range. Example: 10.0.0.0/16,12.34.5.0/24 or 34.56.78.10/32. To build a specific range, you can use this tool: https://www.ipaddressguide.com/cidr. To determine which address is appropriate, contact your IT administrator. |
| Custom DNS Search Suffix | (Optional) If deploying in an existing virtual network with a custom DNS setup, enter the DNS search suffix to use, for example corp.mycompany.internal. This enables headnode and workers to identify themselves using full DNS names of the form: <computer_name>.corp.mycompany.internal. Ensure that the cluster nodes are auto-registered with the DNS service to allow name resolution. If left empty, head node and worker nodes will use their Private IPv4 addresses for communication. |
| Admin Username | Admin username for the cluster. To avoid deployment errors, conform to the list of Disallowed Values (Microsoft) for adminUsername. |
| Admin Password | Choose the password for the admin user of the cluster. This password and the chosen admin username are required to log in into any VM in the cluster using RDP or SSH. For the deployment to succeed, your password must meet Azure's password requirements. See Password requirements when creating a VM for information on the password requirements. |
| License Server | License manager for MATLAB, specified as a string in the form port@hostname. If not specified, online licensing is used. Otherwise, license manager must be accessible from the specified virtual network and subnets. For more information, see https://github.com/mathworks-ref-arch/license-manager-for-matlab-on-azure. |
| MJS Security Level | Security level for the cluster. Level 0: Any user can access any jobs and tasks. Level 1: Accessing other users' jobs and tasks issues a warning. However, all users can still perform all actions. Level 2: Users must enter a password to access their jobs and tasks. The job owner can grant access to other users. |
| Enable Autoscaling | Option indicating whether VM autoscaling is enabled. For more information about autoscaling, refer to the 'Use Autoscaling' section in the deployment README. |
| Automatically Terminate Cluster | Option to autoterminate the cluster after a few hours or when idle. When the cluster is terminated, all worker nodes are deleted and the headnode is deallocated. Select 'Never' to disable auto-termination now but you can enable it later. Select 'Disable auto-termination' to fully disable this feature or if you do not have the permissions to create and assign roles in your subscription. For more information, see Automatically terminate the MATLAB Parallel Server cluster. |
| MJS Scheduling Algorithm | Scheduling algorithm for the job manager. 'standard' spreads communicating jobs across as few worker machines as possible to reduce communication overheads and fills in unused spaces on worker machines with independent jobs. Suitable for good behavior for a wide range of uses including autoscaling. 'loadBalancing' distributes load evenly across the cluster to give as many resources as possible to running jobs and tasks when the cluster is underutilized. |
| Optional User Command | Provide an optional inline PowerShell command to run on machine launch. For example, to set an environment variable CLOUD=AZURE, use this command excluding the angle brackets: <[System.Environment]::SetEnvironmentVariable("CLOUD","AZURE", "Machine");>. You can use either double quotes or two single quotes. To run an external script, use this command excluding the angle brackets: <Invoke-WebRequest "https://www.example.com/script.ps1" -OutFile script.ps1; .\script.ps1>. Find the logs at 'C:\ProgramData\MathWorks\startup.log'. |
| Image ID | Optional Resource ID of a custom managed image in the target region. To use a prebuilt MathWorks image instead, leave this field empty. If you customize the build, for example by removing or modifying the included scripts, this can make the image incompatible with the provided ARM template. To ensure compatibility, modify the ARM template or image accordingly. |
NOTE: If you are using the network license manager, the port and hostname of the network license manager must be reachable from the MATLAB Parallel Server™ virtual machines (VMs). Deploying into a subnet within the same virtual network as the network license manager is a best practice.
-
Click the Review + create button to review the Azure Marketplace terms and conditions.
-
Click the Create button.
When you click the Create button, the resources are created using Azure template deployments. Template deployment can take several minutes.
-
Clicking Create opens the Deployment Details page, where you can monitor the progress of your deployment. Wait for the message Your deployment is complete.
-
Go to your resource group, and select the Storage Account starting with mwstorage. The screen must look like the one in Figure 1.
Figure 1: Resource Group On Completion
-
Under Data Storage on the left panel, click File shares, and select the file share named "shared".
-
Click Browse on the left panel, then open the "cluster" folder.
-
Download the file,
<NAME>.mlsettings, where NAME is the name of your MATLAB job scheduler. -
Open MATLAB.
-
In the Parallel drop-down menu in the MATLAB toolstrip select Create and Manage Clusters.
-
Click Import.
-
Select the downloaded profile and click open.
-
Click Set as Default.
-
(Optional) Validate your cluster by clicking the Validate button.
After you set the cloud cluster as default, the next time you run a parallel language command (such as parfor, spmd, parfeval or batch), MATLAB connects to the cluster. The first time you connect, you are prompted for your MathWorks® account login. The first time you run a task on a worker, it takes several minutes for the worker MATLAB to start. This delay is due to provisioning the instance disk. This is a one-time operation, and subsequent tasks begin faster.
Your cluster is now ready to use.
NOTE: Use the profile and client IP address range to control access to your cloud resources. Anyone with this file can connect to your resources from a machine within the specified IP address range and run jobs on it.
Your cluster remains running after you close MATLAB.
You can remove the resource group and all associated resources when you are done with them. Note that you cannot recover resources once they are deleted. After you delete the cloud resources, you cannot use the downloaded profile again.
-
Log in to the Azure Portal.
-
Select the resource group containing your resources.
-
Select the "Delete resource group" icon to destroy all resources deployed in this group.
-
You are prompted to enter the name of the resource group to confirm the deletion.
To deploy the MATLAB Parallel Server cluster without public IP addresses, set the createPublicIPAddress parameter to No. This prevents creation of public IPv4 addresses for the head node and VMSS worker nodes. The head node and worker nodes use their Private IPv4 addresses to communicate with each other and MATLAB clients. Ensure to meet the following requirements before deploying in a private networking configuration:
Specify the private IPv4 addresses of the MATLAB client(s) that will access the cluster using the clientIPAddressList parameter. This accepts a comma-separated list of IPv4 addresses or CIDR ranges. Example: 10.0.1.0/24,10.0.2.50. You can leave this parameter empty if the MATLAB clients are in the same (or peered) virtual network as the cluster.
If the MATLAB Client is in your on-premises network, ensure that your on-premises network has an existing connection to Azure.
The head node and worker nodes require outbound internet access for:
- Online licensing: Access to
*.mathworks.comfor MATLAB Parallel Server licensing. - Autoscaling and auto-termination.
- Cluster storage account: Access to the storage account created during deployment to upload the cluster profile and other required files. The file share in this account also acts as a shared filesystem for the worker nodes.
When the template creates a new virtual network, the defaultOutboundAccess property is set to true for the default subnet, enabling outbound internet connectivity. If you convert the newly created subnet to a private subnet, you must provision an explicit outbound connectivity method (e.g. NAT Gateway) to allow internet access for the cluster.
If you are using an existing virtual network instead, ensure that it is configured to allow outbound internet access for the head node and worker nodes.
If your virtual network uses Azure Firewall to control outbound traffic, ensure that the following FQDNs are white-listed:
| FQDN | Purpose |
|---|---|
login.microsoftonline.com |
Azure managed identity authentication |
management.azure.com |
Azure Resource Manager API calls (for autoscaling and auto-termination) |
*.blob.core.windows.net |
Azure Blob Storage access |
*.file.core.windows.net |
Azure File Storage access |
*.mathworks.com |
MATLAB online licensing (if applicable) |
See: https://learn.microsoft.com/en-us/azure/firewall/tutorial-firewall-deploy-portal for more information on configuring Azure Firewall rules.
To use a custom DNS server instead of Azure-provided DNS in the Virtual Network containing the cluster, set the customDNSSearchSuffix parameter to your domain's DNS search suffix. If this parameter is left empty, the instances use Private IP addresses instead of hostnames for communication. For MATLAB R2022b and older releases, see the Support for MATLAB release R2022b and older section for additional configuration steps.
Example: With customDNSSearchSuffix parameter set to mycorp.internal.com, worker nodes will communicate with each other and MATLAB Clients using FQDNs of the form <computer_name>.mycorp.internal.com.
Warning: Worker registration with the head node and job submissions will fail if DNS resolution is not properly configured.
Before deploying the cluster:
- Test DNS resolution in your virtual network
- Ensure the private DNS zone linked to the virtual network is configured to automatically register DNS A records for new virtual machines
For more information, see: Configure DNS name resolution for Azure virtual networks
MATLAB R2022b and older releases do not support the usage of private IP addresses for cluster communication. For these releases, we use the computer names of the head node and worker nodes for inter-cluster and client-cluster communication, unless a customDNSSearchSuffix is configured.
When the customDNSSearchSuffix parameter is empty, the cluster uses short hostnames (e.g., parallelserver-headnode) that are not resolvable outside its own network. Your client configuration depends on its location.
- MATLAB Client in the Same Virtual Network: No extra configuration is needed. The MATLAB client can resolve the cluster's short hostnames using Azure's default DNS.
- MATLAB Client in a Peered Virtual Network: You must enable name resolution across the networks so that the MATLAB client can resolve hostnames of the cluster nodes. This can be done using Azure Private Zones:
- Create an Azure Private DNS zone.
- Link the zone to both the cluster's and the client's virtual networks, ensuring auto-registration is enabled.
- Set the
customDNSSearchSuffixparameter to the name of your private DNS zone.
Before you can use your MATLAB Parallel Server cluster, you must configure certain required ports on the cluster and client firewall. These ports allow your client machine to connect to the cluster headnode and facilitate communication between the cluster nodes.
For details about the port requirements for cluster nodes, see this information from MathWorks® Support Team on MATLAB Answers: How do I configure MATLAB Parallel Server using the MATLAB Job Scheduler to work within a firewall?.
Additionally, if your client machine is outside the cluster’s network, then you must configure the network security group of your cluster to allow incoming traffic from your client machine on the following ports. For information on how to configure your network security group, see Create a security rule. To troubleshoot, see this page.
| Required ports | Description |
|---|---|
| TCP 27350 to 27358 + 4*N | For connecting to the job manager on the cluster headnode and to the worker nodes for parallel pools. Calculate the required ports based on N, the maximum number of workers on any single node across the entire cluster. |
| TCP 443 | If you are using online licensing, you must open this port for outbound communication from all cluster machines. If you’re using Network License Manager instead, then you must configure ports as listed on Network License Manager for MATLAB on Microsoft Azure. |
| TCP 3389 | Remote Desktop Protocol to access to cluster nodes. |
| TCP 22 | SSH access to cluster nodes. |
Table 1: Outgoing port requirements
By default, MATLAB Parallel Server is configured with the public hostname of each machine to allow the MATLAB client to access both the scheduler and workers. If you modify the Azure Resource Manager (ARM) template provided, make sure that you provide a public hostname for the head node and the worker nodes.
To optimize the number of virtual machines running MATLAB workers, enable autoscaling by setting Enable Autoscaling to Yes when you deploy the template. Autoscaling is optional and is disabled by default.
When autoscaling is disabled, the Virtual Machine Scale Set (VMSS) deploys Num Worker Nodes instances. To change the number of worker nodes, use the Azure Portal.
If you enable autoscaling, the capacity of the VMSS is regulated by the number of workers needed by the cluster. The number of virtual machines is initially set to Num Worker Nodes. It then fluctuates between Min and Max Worker Nodes. To change these limits after deployment, use the Azure Portal and modify the tags on the VMSS. To change the amount of time for which idle nodes are preserved, adjust the value of the tag mwWorkerIdleTimeoutMinutes. Do not use the "manual scale" option to change the instance count, as doing so can lead to the unsafe termination of MATLAB workers.
Ensure that the Max Worker Nodes parameter is within your Azure subscription quotas for the specific instance type. To learn about setting quotas, see Azure subscription and service limits, quotas, and constraints.
To disable autoscaling after the deployment, redeploy the template with autoscaling disabled.
Use the Automatically Terminate Cluster parameter while deploying the resource group to manage costs efficiently. You can choose one of these options:
Never(default): No auto-termination but can be enabled after deployment.When cluster is idle: Terminates the cluster when it is idle for about 10 minutes (30 minutes at startup).After x hours: Terminates the cluster after 'x' hours (wherexis between 1 and 24).Disable auto-termination: No auto-termination. Use this option to fully disable this feature or if you do not have the permissions to create and assign roles in your subscription.
When the cluster is auto-terminated, the head node virtual machine is deallocated and all worker virtual machines are deleted. To use the cluster again, restart the head node.
To modify the termination policy after deploying the resource group, edit the value of the tag mw-autoshutdown that is attached to the head node. Set the value of the tag to either never, on_idle, or After x hours, where x must be an integer between 1 and 24.
By default, MATLAB Job Scheduler (MJS) is configured to manage a wide range of cluster uses.
To change the MJS configuration for advanced use cases, replace the default mjs_def file with your own file using the template parameter OptionalUserCommand. Doing so overwrites all MJS startup parameters, except for DEFAULT_JOB_MANAGER_NAME, HOSTNAME, and SHARED_SECRET_FILE. For more information about the MJS startup parameters and to edit them, see Define MATLAB Job Scheduler Startup Parameters.
For example, to retrieve and use your edited mjs_def from a storage service (such as Azure blob storage), set the OptionalUserCommand to the following:
Invoke-WebRequest "https://<your_storage_account>.blob.core.windows.net/<container_name>/mjs_def.bat" -OutFile $Env:MJSDefFile
If your resource group fails to deploy, check the Deployments section of the Resource Group. This section indicates which resource deployments failed and allow you to navigate to the causing error message.
If the resource group deployed successfully but you are unable to validate the cluster, check the logs on the instances to diagnose the error. The deployment logs are output to C:\ProgramData\MathWorks\ on the instance nodes. The cluster logs are output to C:\ProgramData\MJS\ and the system startup logs are output to C:\ProgramData\MathWorks\startup.log on every instance of the cluster.
If your cluster is unable to create new workers, you encounter scaling errors. To troubleshoot these issues, see Resolve Scaling Issues in the Cloud.
Copyright 2018-2026 The MathWorks, Inc.