You must create a cluster before you can use E-HPC. This guide walks you through quickly creating an E-HPC cluster in the console.
Prerequisites
-
The E-HPC service-linked role is created. The system prompts you to create this role when you first log on to the E-HPC console.
-
A VPC and vSwitch are created. Create a VPC and Create a vSwitch.
-
Storage resources are created. E-HPC clusters support NAS and CPFS file systems.
-
NAS: Activate the NAS service, then Create a file system and Add a mount target.
-
CPFS-POSIX: Activate the CPFS service, then Create a file system and Manage POSIX mount targets.
-
CPFS-NFS: Activate the CPFS service, then Create a file system, Manage protocol services, and Manage export directories.
-
Procedure
-
Log on to the E-HPC console.
-
In the upper-left corner of the top navigation bar, select a region.
-
In the left-side navigation pane, click Cluster.
-
On the Cluster page, click Create Cluster.
NoteThis guide shows how to quickly create a cluster with a basic configuration. For parameters not covered here, keep the default settings. For more information, see Create a cluster by using the wizard.
-
On the Create Cluster page, configure the hardware settings.
Parameter
Example
Description
Zone
China (Hangzhou) Zone I
The availability zone for the cluster.
NoteTo ensure efficient network communication between nodes, all nodes in an E-HPC cluster must be deployed in the same region and availability zone. For more information, see Regions and availability zones.
Pricing Model
Pay-as-you-go
The billing method for cluster nodes. This example uses pay-as-you-go, where you are billed hourly based on actual usage.
For more information about billing methods, see Billing overview.
Deploy Mode
Standard
The deployment mode for the cluster. This example uses Standard mode, which deploys login, management, and compute nodes on separate instances.
Node type and quantity
-
Compute Node: ecs.gn5-c4g1.xlarge
-
Management Node: ecs.c7.large
-
Login Node: ecs.c7.large
The instance types and number of nodes. For guidance on selecting instance types, see Product specifications and Instance Type Selection Best Practices.
-
Compute Node: A node that runs HPC jobs. Compute node specifications determine the overall performance of the E-HPC cluster.
-
Management Node: A node that handles job scheduling and domain account management.
-
Login Node: E-HPC assigns an elastic IP address to this node by default. You can remotely log on to this node to manage the cluster from the command line.
-
-
Expand the Advanced Configurations section to configure network, storage, and other parameters.
Configuration area
Parameter
Example
Description
Network Configuration
VPC
vpc-uf6iprpvifxp5i6******
The VPC for the cluster. VPCs are logically isolated from each other. The system selects the first VPC and vSwitch you created by default. Make sure the vSwitch has enough available IP addresses for all nodes in the cluster.
vSwitch
vsw-uf643l4q2bk06wz******
Storage
File System ID
353cc4**** (Performance NFS)
The system selects the first file system ID and mount target you created by default. Make sure the file system has enough available mount targets.
Mount Point
353cc4****-ek***.cn-shanghai.nas.aliyuncs.com
-
Click Next and configure software settings.
Parameter
Example
Description
Image Type
public image
The image type determines the operating system.
Image
CentOS_7.6_64
When the cluster is created, the system copies the OS and application data from the selected image to all cluster nodes.
Scheduler
pbs
The job scheduler for the E-HPC cluster.
Domain Account Service
nis
The user and account management service for the cluster.
VNC
Enable VNC
Enabling VNC automatically deploys a remote visualization service during cluster creation. You can then access a remote graphical desktop directly from the console in a web browser.
-
Click Next and complete the basic settings.
Parameter
Example
Description
Name
test
The cluster name, which appears on the cluster list page.
Login Password and Repeat Password
Ehpc123***
The cluster password. The
rootuser uses this password to remotely log on to the login node over SSH. -
In the Configuration List pane on the right, review your cluster configuration. Read and select Alibaba Cloud International Website Product Terms of Service, and then click OK.
Result
After the cluster is created, go to the Clusters page to check its status. The cluster is ready when all nodes are in the Running state.