View, Edit, Start, or Stop Clusters
View and Start Clusters
To see a list of your MATLAB® Parallel Server™ clusters, in Cloud Center click Cloud Resources.
Your cluster can be in one of three states.
Running: If you start a cluster in the cloud, the cluster machines and data volumes are running and the cluster has an IP address. The data volumes consist of the MATLAB Job Scheduler data and any enabled shared storage.
Stopped and Online: If your running cluster reaches the end of its Auto-Shutdown period—whether set to a specific number of hours, On Idle (about 10 minutes of inactivity, or 30 minutes at startup), or Always On—the cluster machines stop, but the data volumes remain active and continue to incur costs. These resources maintain your existing cluster access rules and allow you to restart the cluster quickly. To avoid costs, click Stop All Resources in the warning message to stop these resources and save a data backup of the cluster as an Amazon® Elastic Block Store (EBS) snapshot in your Amazon Web Services (AWS®) account.
Stopped and Offline: If you save a cluster without starting it, or if you stop a running cluster, neither the cluster machines nor data volumes are running. When you stop a cluster, a new backup of your data volumes is saved as an EBS snapshot in your AWS account. Restarting the cluster does not delete the backup from your AWS account. You must manually delete old backups from your AWS account if you no longer need them.
| State of Cluster | Cluster Machines | Data Volumes | New Data Backup |
|---|---|---|---|
| Running | ✔ | ✔ | ❌ |
| Stopped and Online | ❌ | ✔ | ❌ |
| Stopped and Offline | ❌ | ❌ | ✔ |
The following listing shows a pair of clusters, one currently Running, the other Stopped and Offline:
To start up your clusters from the Stopped states, click Start in the Actions column.
To access a running cluster created in your account, use Discover Clusters from MATLAB. See Discover Clusters. For detailed information about a particular cluster, click its name in the list.
Stop or Delete Clusters
To stop a running cluster, click Stop in the Actions column. If your cluster has timed out after the hours you set in Auto-Shutdown (hrs) period, you can stop the data volumes in your cluster by clicking Stop All Resources in the warning message. Stopping a cluster does not remove it from your list. You can start the cluster again later.
Stop removes all resources and saves a backup of your data volumes as an EBS snapshot in your AWS account. Each time you stop a cluster, a new snapshot is created in your AWS account.
Your cluster can receive updates after restart, and you can change the machine type and other cluster attributes that you cannot change while the cluster is running. These updates include any latest MATLAB updates and operating system updates.
You do not need to reimport the cluster profile into MATLAB after restarting the cluster. For sharable clusters, you do not need to reshare the cluster profile.
You can stop a cluster during its startup if it fails to start, takes too long, or you change your mind.
To permanently remove a cluster, click Delete. When you delete a cluster, it will be removed from Cloud Center but a data backup of the MATLAB Job Scheduler data is saved as an EBS snapshot to your cloud account.
Edit Cluster
You can select Edit under the three dots in the Actions column to change the following attributes of the cluster depending on whether your cluster is online or offline.
Online: If you start a cluster in the cloud, the cluster is online and has an IP address. If your running cluster reaches the end of its Auto-Shutdown (hrs) period, the cluster remains online because the data volumes are still running.
Offline: If you have saved a cluster without starting it or if you have stopped a running cluster, the cluster becomes offline.
Edit Cluster Attributes
| Attribute | Online | Offline |
|---|---|---|
| Name | ✔ | ✔ |
| Shared State | ❌ | ❌ |
| Auto-Manage Cluster Access | ✔ | ✔ |
| Headnode Machine Type | ❌ | ✔ |
| Worker Machine Type | ❌ | ✔ |
| Allow cluster to auto-resize | ❌ | ✔ |
| Upper Limit | ❌ | ✔ |
| Worker Per Machine | ❌ | ✔ |
| Auto-Shutdown (hrs) | ❌ | ✔ |
| Log Level | ✔ | ✔ |
| VPC | ❌ | ✔ |
| Subnet | ❌ | ✔ |
| Enable Shared Storage | ❌ | ✔ |
| Storage Size (GB) | ❌ | ✔ |
| Name of SSH Key | ❌ | ✔ |
| Default IP Address | ❌ | ✔ |
Click Save to save your changes.
Share Cluster
You can download your cluster profile and share a cluster only if you have a shared cluster. If you have a shared cluster, you can share your cluster with others by giving them explicit access using the cluster profile (downloaded from the Actions column in Cloud Center or exported from MATLAB). Anyone who has the cluster profile can import the cluster profile into MATLAB, allowing them to submit jobs to or interact with the cluster from MATLAB. They can also reshare access to the cluster by resharing the cluster profile.
Copy Cluster
You can select Duplicate under the three dots in the Actions column to copy your cluster.
To share the configuration settings of your cluster, select Copy Resource Configuration Settings under the three dots in the Actions column. This action copies the resource configuration settings of your cluster as a URL to your clipboard that you can share with other users.
