Create a customer-managed fleet

Open the Deadline Cloud console.

Select Farms. A list of available farms displays.

Select the name of the Farm you want to work in.

Select the Fleets tab.

Choose Create fleet.

Enter a Name for your fleet.

(Optional) Enter a Description for your fleet.

Select Customer managed for Fleet type.

Select an Auto Scaling type. For more information, see Use EventBridge to handle Auto Scaling events.

Select your fleet's service access.

1. We recommend using the Create and use an new service role option for each fleet for more granular permissions control. This option is selected by default.

2. You can also use an existing service role by selecting Choose a service role.

Review your selections, then choose Next.

Select an operating system for your fleet. All of a fleet’s workers must have a common operating system.

Select the host CPU architecture.

Select the minimum and maximum vCPU and memory Hardware capabilities to meet the workload demands of your fleets.

(Optional) Select the arrow to expand the Add capabilities section.

(Optional) Select the checkbox for Add GPU capability - Optional, then enter the minimum and maximum GPUs and memory.

Review your selections, then choose Next.

(Optional) Define custom worker capabilities, then choose Next.

Using the dropdown, select one or more queues to associate with the fleet.

Review the queue associations, then choose Next.

(Optional) For Default Conda queue environment, we'll create an environment for your queue that will install Conda packages requested by jobs.

(Optional) Add tags to your CMF. For more information, see Tagging your AWS resources.

Review your fleet configuration and make any changes.

Choose Create fleet.

Select the Fleets tab, then note the Fleet ID.

Open a terminal.

Create fleet-trust-policy.json in a new editor.

1. Add the following IAM policy, replacing the ITALICIZED text with your AWS account ID and Deadline Cloud farm ID.

   {
       "Version": "2012-10-17",
       "Statement": [
           {
               "Effect": "Allow",
               "Principal": {
                   "Service": "credentials.deadline.amazonaws.com"
               },
               "Action": "sts:AssumeRole",
               "Condition": {
                   "StringEquals": {
                       "aws:SourceAccount": "ACCOUNT_ID"
                   },
                   "ArnEquals": {
                       "aws:SourceArn": "arn:aws:deadline:*:ACCOUNT_ID:farm/FARM_ID"
                   }
               }
           }
       ]
   }

2. Save your changes.

Create fleet-policy.json.

1. Add the following IAM policy.

   {
       "Version": "2012-10-17",
       "Statement": [
           {
               "Effect": "Allow",
               "Action": [
                   "deadline:AssumeFleetRoleForWorker",
                   "deadline:UpdateWorker",
                   "deadline:DeleteWorker",
                   "deadline:UpdateWorkerSchedule",
                   "deadline:BatchGetJobEntity",
                   "deadline:AssumeQueueRoleForWorker"
               ],
               "Resource": "*",
               "Condition": {
                   "StringEquals": {
                       "aws:PrincipalAccount": "${aws:ResourceAccount}"
                   }
               }
           },
           {
               "Effect": "Allow",
               "Action": [
                   "logs:CreateLogStream"
               ],
               "Resource": "arn:aws:logs:*:*:*:/aws/deadline/*",
               "Condition": {
                   "StringEquals": {
                       "aws:PrincipalAccount": "${aws:ResourceAccount}"
                   }
               }
           },
           {
               "Effect": "Allow",
               "Action": [
                   "logs:PutLogEvents",
                   "logs:GetLogEvents"
               ],
               "Resource": "arn:aws:logs:*:*:*:/aws/deadline/*",
               "Condition": {
                   "StringEquals": {
                       "aws:PrincipalAccount": "${aws:ResourceAccount}"
                   }
               }
           }
       ]
   }

2. Save your changes.

Add an IAM role for the workers in your fleet to use.

aws iam create-role --role-name FleetWorkerRoleName --assume-role-policy-document file://fleet-trust-policy.json
aws iam put-role-policy --role-name FleetWorkerRoleName --policy-name FleetWorkerPolicy --policy-document file://fleet-policy.json

Create create-fleet-request.json.

1. Add the following IAM policy, replacing the ITALICIZED text with your CMF's values.

   Note

   You can find the ROLE_ARN in the create-cmf-fleet.json.

   For the OS_FAMILY, you must choose one of linux, macos or windows.

   {
       "farmId": "FARM_ID",
       "displayName": "FLEET_NAME",
       "description": "FLEET_DESCRIPTION",
       "roleArn": "ROLE_ARN",
       "minWorkerCount": 0,
       "maxWorkerCount": 10,
       "configuration": {
           "customerManaged": {
               "mode": "NO_SCALING",
               "workerCapabilities": {
                   "vCpuCount": {
                       "min": 1,
                       "max": 4
                   },
                   "memoryMiB": {
                       "min": 1024,
                       "max": 4096
                   },
                   "osFamily": "OS_FAMILY",
                   "cpuArchitectureType": "x86_64",
               },
           },
       }
   }

2. Save your changes.

Create your fleet.

aws deadline create-fleet --cli-input-json file://create-fleet-request.json