Arctera Application Mobility Service Help

Last Published:
Product(s): InfoScale & Storage Foundation (8.0.2, 7.4.2, 7.4.1, 1.0)
Platform: Linux

Cluster deployment

The Application Mobility Service allows users to deploy a new resilient cluster to AWS as well as Azure platform. While the basic mechanism remains the same, there are a few jargon-specific differences. The following table depicts the terminology comparison for AWS and Azure cluster deployment.

 

AWS

Azure

Resiliency

Zones and subnets are interdependent

Zones and subnets are independent

Selected resource group info

Not available

Available

Compute section

EC2 type

Virtual machine type

Storage and file system

gp2/gp3

Standard_LRS, etc.

To deploy a new cluster in the AWS or Azure cloud, perform the following steps:

  1. Click Create deployment plan.

    Note:

    Mandatory fields are indicated by an asterisk (*) sign.

  2. Under Basic details, provide a Deployment plan name with character count limit between 4 and 256. Select a target datacenter from the list of 'Connected' datacenters for the deployment, and choose Application either for General use (default) or Oracle.

    Note:

    The application type cannot be changed for the plan later. In that case, create a new plan and start over the process.

    Click Save and exit for later use. Otherwise, click Next to continue.

    A message 'Deployment plan name created successfully' is displayed.

  3. The configuration stage requires several parameters to be defined as follows:

    • Resource prefix: The Application Mobility Service prepends the resource prefix name to all cloud resources created through deployment plan execution. You can then use this resource prefix to search all cloud resources in the AWS or Azure portal.

      Ensure that you meet the following conditions while assigning the resource prefix name:

      • It should be unique for every plan. Else, migration plan execution fails.

      • The maximum character length allowed is 24.

      • Permissible characters for AWS plan include only alphanumeric characters and hyphens (-).

      • Permissible characters for Azure plan include only lowercase letters, numerals, and hyphens (-).

    • Resiliency: Select the level of resiliency required for the deployment.

      • Select Within availability zone or Across availability zone as per the requirement.

      • First availability zone: Select the primary availability zone from the dropdown list for your datacenter.

      • Second availability zone: In the case of across availability zone, select the secondary availability zone from the list.

        In the Azure case, if one AZ is selected under 'First Availability Zone', then the remaining AZs are shown under 'Second Availability Zone'.

      • Select subnets: Existing

        For Availability zones, select the required Eth0 and Eth1 subnet names per availability zone, whereas the Eth2 field is optional.

        Note:

        Since Azure zones and subnets are independent, you can select same subnets for both the Eth0 cases.

      • For the Azure case, a read-only Selected resource group field is available. The information is fetched from the gateway registration process.

    • Compute: Use default parameters or provide Cluster Size, EC2 type or Virtual machine type, and Security key as per the target performance requirements.

      Note:

      For Security key, ensure that .pem and .pub files are placed in the same directory, and you provide the path of .pem file here. Otherwise, New deployment plan creation fails throwing an undefined error. For more information, refer Prerequisites subsection of the Introduction chapter.

      Select Operating system as Redhat8 or RedHat9.

      Select InfoScale Version as 8.0 or 8.0.2.

      To exercise the Custom Machine Image feature, select the box next to Bring your own image and provide the correct Machine Image ID or name.

      Note:

      The bandwidth and latency of the network influences the time taken to execute a Custom Machine Image plan. Therefore, if timeouts occur and Download RPM task fails, retry it a few times to proceed.

    • Storage and file system: This option lets you customize storage capacity needed for your application.

      Default mount paths are already defined and listed. You can modify them, remove them, or add new mount paths as required.

      Ensure that the mount point size should be any non-zero positive value.

      If Configure storage is set to No, all the default mount paths are removed, and the Review page displays the Mount Paths value equal to "0".

      Note:

      If Configure storage is set to Yes, Application Mobility configures SFCFSHA (Storage Foundation Cluster File System High Availability) on the newly created cluster; otherwise, it configures only VCS on the newly created cluster.

    • Application: Provide Oracle parameters such as SID, Owner, Home, Application IP, Port, and Protocol to continue.

      Note:

      This section is active only when Application type is selected as Oracle in the Basic details step.

    • Fencing: This Application Mobility Service feature supports Coordination Point (CP) server-based fencing for cluster deployment. It allows you to deploy CP servers and configure them in cluster servers to prevent data corruption.

      For more information, refer Cluster Server Administrator's Guide-Linux.

      Note:

      Fencing mode is supported for all the application types and the cloud platforms (AWS or Azure).

      Under Fencing, click Select Fencing Mode. There are three fencing modes available:

      • Disabled:

        • This is the default selection.

        • With this selection, the Application Mobility Service does not deploy CP servers.

      • CPS-Based (1 Node):

        • Provides the option to deploy a single CP server in the selected availability zone.

        • Choose one of the following deployment options:

          • New: Select subnets from the available options.

          • Existing: Select a CP server from the available options.

      • CPS-Based (3 Node):

        • Provides the option to deploy a three-node CP server cluster across the selected availability zones.

        • Choose one of the following deployment options:

          • New: Select subnets from the available options.

          • Existing: Select CP servers from the available options.

    • You can Cancel the operation or Save and exit for later use, or go to Previous step at any stage of the process. Click Next to continue.

  4. The Review stage provides the summary of all the inputs where you can verify or edit them, if required. Click Edit to go to respective section and update them.

    Click Execute to complete the deployment process. A message "Plan execution for plan name started successfully" is displayed.

  5. Click Refresh to update the deployment plan list. You can check the overall progress of the operation by clicking the field under 'Status' column against the plan name.

    Note:

    For the Oracle use case, after successfully executing the new deployment plan, you need to install and configure the Oracle application database instances on the cluster on your own.

  6. Upon cancellation of failed plan, all the associated resources with deployment plan will be cleaned up.
Post Deployment

After deployment is successful, CFSMount service group can be brought to the ONLINE state using the following commands:

[azureuser@azure-custom-ami-az1-0 ~]$ sudo /opt/VRTSvcs/bin/hagrp -online CFSMount -any

VCS NOTICE V-16-1-50735 Attempting to online group on system azure-custom-ami-az1-0

VCS NOTICE V-16-1-50735 Attempting to online group on system azure-custom-ami-az2-0