Help Center/ ModelArts/ ModelArts User Guide (Lite Server)/ Enabling Lite Server Resources
Updated on 2024-11-19 GMT+08:00

Enabling Lite Server Resources

Figure 1 Flowchart for enabling resources

Table 1 Enabling resources

Phase

Task

Preparations

1. Apply for resource specifications.

2. Increase the resource quota.

3. Enable basic permissions.

4. Configure an agency authorization for ModelArts.

5. Create a VPC.

6. (Optional) Create a key pair. (This step is not required if password login is used.)

Purchasing Lite Server resources

7. Purchase a resource pool on the ModelArts console.

Step 1: Applying for Resource Specifications

Contact Huawei Cloud customer manager to determine a resource solution for Lite Server. Apply for the required resource specifications. If there is no customer manager, submit a service ticket.

Step 2: Increasing Resource Quota

The resources required by the server may exceed the default resources (such as ECS, EIP, and SFS) provided by Huawei Cloud. In this case, you need to increase the resource quota.

  1. Log in to Huawei Cloud console.
  2. Hover over Resources from the top navigation bar and choose My Quotas.
  3. Click Increase Quota in the upper right corner, fill in the materials, and submit a service ticket.

    Increase the quota before purchasing and provisioning the resource, ensuring it exceeds the resource's requirements.

Step 3: Enabling Basic Permissions

To enable basic permissions, log in to the management console as the administrator account and assign the basic permissions (such as ModelArts FullAccess, BMS FullAccess, ECS FullAccess, VPC FullAccess, VPC Administrator, and VPCEndpoint Administrator) required by server to IAM users.

  1. Log in to the IAM console.
  2. In the navigation pane on the left, choose User Groups, and then click Create User Group in the upper right corner.
  3. Enter the user group name and click OK.
  4. Click Manage User in the Operation column and add the target users.
  5. Click the user group name to go to the group details page.
  6. In the Permissions tab, click Authorize.

    Figure 2 Assigning permissions

  7. Enter ModelArts FullAccess in the search box and select ModelArts FullAccess.

    Figure 3 ModelArts FullAccess

    Use the same method to select BMS FullAccess, ECS FullAccess, VPC FullAccess, VPC Administrator, and VPCEndpoint Administrator. Server Administrator and DNS Administrator are dependent policies and are automatically selected.

  8. Click Next and set Scope to All resources.
  9. Click OK.

Step 4: Configuring an Agency Authorization for ModelArts

ModelArts accesses other dependent services. Therefore, you need to configure an agency authorization for ModelArts. For details, see Assigning Permissions to Individual Users for Using ModelArts.

Step 5: Creating a VPC

To create a VPC, you need to log in to the management console as the administrator.

  1. Log in to the management console.
  2. In the service list on the left, choose Networking > Virtual Private Cloud.
  3. Click Create VPC in the upper right corner, configure the parameters, and click Create Now. For details about the parameters, see Creating a VPC and Subnet.

    Figure 4 Creating a VPC

Step 6: Creating a Key Pair

Key pairs are not required for password login.

  1. Log in to the ModelArts console.
  2. In the navigation pane on the left, choose AI Dedicated Resource Pools > Elastic Servers.
  3. Click Buy Dedicated AI Server.
  4. Click Create Key Pair.
  5. On the displayed page, select I have read and agree to Key Pair Service Disclaimer, and click OK. Then, save the key pair to the local host.

Step 7: Purchasing Resources

Mapping Between Compute Resources and Image Versions lists the supported BMS images.

  1. Log in to the ModelArts console.
  2. In the navigation pane on the left, choose AI Dedicated Resource Pools > Elastic Servers.
  3. Click Buy Dedicated AI Server and configure the parameters.

    Table 2 Basic parameters

    Parameter

    Description

    Resource Type

    • BMS: A BMS features both the scalability of Elastic Cloud Servers (ECSs) and high performance of physical servers, providing dedicated servers on the cloud for you and your enterprise.
    • ECS: ECS provides secure, scalable, on-demand compute resources, enabling you to flexibly deploy applications and workloads.

    Billing Mode

    Select Pay-per-use or Yearly/Monthly.

    • Yearly/Monthly

      Yearly/Monthly is a prepaid billing mode. You pay in advance for a subscription term, and in exchange, you get a discounted rate. Yearly/Monthly billing is a good option for long-term, stable services.

    • Pay-per-use

      Pay-per-use is a postpaid billing mode. You are charged for how long you use each ECS. You can purchase or delete such an ECS at any time.

    Region

    Select a region near you to ensure the lowest latency possible. After the resources are purchased, you can switch the region in the upper left corner of the console to view the resources.

    AZ

    A standalone data center with an independent network and power supply. When deploying resources, consider your applications' requirements on disaster recovery (DR) and network latency.

    • For high DR capability, deploy resources in different AZs within the same region.
    • For lower network latency, deploy resources in the same AZ.

    If CloudPond is used, you can view the corresponding edge AZ. Edge AZs deploy cloud infrastructure and services at customer premises. In scenarios where there are high requirements on application access latency, local data retention, and local system interaction, edge AZs ensure easy deployment to the local environment. For details about CloudPond, see What Is CloudPond?

    Table 3 Parameters for configuring specifications

    Parameter

    Description

    Server Name

    Server name, which can contain 1 to 64 characters, including letters, digits, hyphens (-), and underscores (_).

    CPU Architecture

    CPU architecture of the resource type, which can be x86 or Arm. Select the CPU architecture based on the required flavors. If GPUs are used, select x86. If NPUs are used, select Arm. The flavors vary by region. The actual flavors are displayed on the console.

    NOTE:

    If no flavor is available on the page, contact Huawei Cloud technical support to apply for the flavor.

    System Disk

    This parameter is displayed only when you select a flavor that supports mounting. After an ECS is created, you can mount a data disk to the ECS or expand the capacity of the system disk on the ECS. The recommended value is not smaller than 100 GB.

    Table 4 Image parameters

    Parameter

    Description

    Image

    • Public image

      A public image is a standard OS image provided by the system and is available to all users. It contains an OS and pre-installed public applications, such as the SDI iNIC driver, bms-network-config (a network configuration program), and Cloud-Init (an initialization tool). If you need other applications or software, configure them on the new BMSs. ModelArts provides images, which support multiple OSs, built-in drivers and software for AI scenarios, and preset custom ModelArts OS optimization components. For details about supported images, see Mapping Between Compute Resources and Image Versions.

    • Private image

      A private image is created from an external image file or a BMS and is available only to the user who created it. It contains an OS, preinstalled public applications, and the user's personal applications. You can select a private image to save your time from repeatedly configuring servers.

    Table 5 Network parameters

    Parameter

    Description

    VPC

    The VPC should be the same as that of other cloud services, such as MapReduce Service (MRS) and Cloud Container Engine (CCE) for network interaction.

    Subnet

    Select a subnet of the current VPC.

    IPv6 Network

    IPv6 is available when it is supported by the current subnet, flavors, and images.

    Ensure that IPv6 has been enabled. For details about enabling IPv6, see Creating a Subnet for the VPC.

    The requirements on flavors and images differ. Refer to the information displayed on the console.

    RoCE Network

    When GPUs of series A are used, you need to configure the RoCE network to use the RoCE NICs on the hardware during distributed training.

    This parameter is related to the selected flavors. RoCE network will not be available if any flavor does not support RoCE.

    If RoCE is supported but has not been created, click Create RoCE.

    If RoCE is supported and has been created, select an existing one. Repeated creation is not supported.

    Security Group

    A security group is a collection of access control rules for ECSs that have the same security requirements and that are mutually trusted within a VPC.

    Table 6 Management parameters

    Parameter

    Description

    Login Mode

    Key pair is recommended because it features higher security than Password. If you select Password, ensure that the password meets complexity requirements to prevent malicious attacks.

    • Key pair

      A key pair is used for BMS login authentication. You can select an existing key pair, or click Create Key Pair to create one.

      NOTE:

      If you use an existing key pair, ensure that you have one.

    • Password

      The initial password is used for authentication. You can log in to the BMS using the username and its initial password.

      If the BMS runs Linux, you can use username root and its initial password to log in to the BMS. If the BMS runs Windows, you can use username Administrator and its initial password to log in to the BMS. Password complexity must meet the following requirements:

      • Contains 8 to 26 characters.
      • Contains at least three types of the following characters: uppercase letters, lowercase letters, digits, and special characters (!@$ %^-_=+[{}]:, ./?).
      • The password should be different from the username or the username spelled backwards.
      • Cannot contain root, administrator, or their reverse.
    Table 7 Advanced settings

    Parameter

    Description

    Enterprise Project

    This parameter is available only if you have enabled enterprise projects or your account is an enterprise account. You can contact your service manager to enable this function

    An enterprise project groups cloud resources, so you can manage resources and members by project. The default project is default.

    Select an enterprise project from the drop-down list. For details about enterprise projects, see Enterprise Management User Guide.

    Table 8 Purchase parameters

    Parameter

    Description

    Quantity

    You can purchase multiple instances simultaneously, with a value between 1 and 10.

    Each instance generates a separate order, which must be paid for individually.

  4. Click Create Now.
  5. Pay for the order.

    Each instance generates a separate order, which must be paid for individually.

  6. Once paid, the resource will be created in 20 to 60 minutes. Wait until the resource is created.

    • If ModelArts fails to create an elastic server, there are multiple possible causes. The following provides several possibilities for quick troubleshooting.
      • Insufficient resources: Switch to the BMS page and check whether the specifications to be purchased are sold out. If so, there are no resources of this flavor. In this case, contact the customer manager to obtain resources and purchase again.
      • Insufficient quota: Check whether the resource quota of the account is sufficient. If the resource quota of the account, including the number of cores and RAM, is insufficient, the creation will fail. In this case, apply for a quota before purchasing the resource.
      • Internal BMS error: Check whether there is an internal BMS error. If yes, submit a service ticket to BMS to locate and rectify the fault.
    • If a container is used or shared by multiple users, you should restrict the container from accessing the OpenStack management address (169.254.169.254) to prevent host machine metadata acquisition. For details, see Forbidding Containers to Obtain Host Machine Metadata.