question re having multiple containers in a task definition

[Gerald DeConto]

I have a situation where we want to process data from 3 different queues, and the processing can be done by the same code (container).

The AWS documentation says I can define multiple containers in a single task definition but I am having problems setting up this scenario

My tests (see code snippets below) setup returns the following error (where instance_prefix='mycode') when I deploy:

    * aws_ecs_service.ecs_service: InvalidParameterException: The container mycode does not exist in the task definition.

Which makes sense since the containers are named mycode0, mycode1 and mycode2.

Question1: any thoughts on how to fix this so that the deploy works?

Question2: would it be better to define three separate tasks instead?  one for mycode0, one for mycode1, etc

Any help appreciated.

# part of the main.tf

resource "aws_ecs_task_definition" "mytask" {

  family = "${var.instance_prefix}"

  container_definitions = "${template_file.task_def.rendered}"

}

module "cluster" {

  source  = "../modules/ecs-cluster"

  cluster_name = "${var.instance_prefix}"

  environment = "${var.environment}"

  aws_ami = "${var.aws_ami}"

  key_name = "${var.key_name}"

  aws_security_group = "${var.aws_security_group}"

  aws_vpc_subnets = "${var.aws_vpc_subnets}"

  instance_type = "${var.instance_type}"

  asg_max_size = "${var.asg_max_size}"

  asg_min_size = "${var.asg_min_size}"

  asg_desired_capacity = "${var.asg_desired_capacity}"

}

module "service" {

  source = "../modules/ecs-service"

  cluster_name = "${module.cluster.name}"

  service_name = "${var.instance_prefix}"

  task_definition = "${aws_ecs_task_definition.collector.arn}"

  task_count = "${var.instance_count}"

  container_name = "${var.container_name}"

  container_port = "${var.container_port}"

  elb_id = "${aws_elb.collector.id}"

}

# part of the module that defines the ecs_service:

resource "aws_ecs_service" "ecs_service" {

  name = "${var.service_name}"

  cluster = "${var.cluster_name}"

  task_definition = "${var.task_definition}"

  desired_count = "${var.task_count}"

  iam_role = "${aws_iam_role.ecs_service.arn}"

  load_balancer {

    elb_name = "${var.elb_id}"

    container_name = "${var.container_name}"

    container_port = "${var.container_port}"

  }

  # Workaround for "timeout while waiting for state to become 'INACTIVE'" error frequently seen during destroy.

  # See https://github.com/hashicorp/terraform/issues/2902. Seems that the service cannot be removed if the IAM

  # role policy is destroyed before the service.

  depends_on = ["aws_iam_role_policy.ecs_service"]

}

# task definition template file snippet:

[

  {

    "cpu": 50,

    "command": [

      "mycode.py"

    ],

    "environment": [

      {

        "name": "MYVAR",

        "value": "0"

      }

    ],

    "essential": true,

    "image": "docker.moveaws.com:443/sdsread/mycode:latest",

    "memory": 500,

    "name": "mycode0"

  },

  {

    "cpu": 50,

    "command": [

      "mycode.py"

    ],

    "environment": [

      {

        "name": "MYVAR",

        "value": "1"

      }

    ],

    "essential": true,

    "image": "docker.moveaws.com:443/sdsread/mycode",

    "memory": 500,

    "name": "mycode1"

  },

  {

    "cpu": 50,

    "command": [

      "mycode.py"

    ],

    "environment": [

      {

        "name": "MYVAR",

        "value": "2"

      }

    ],

    "essential": true,

    "image": "docker.moveaws.com:443/sdsread/mycode",

    "memory": 500,

    "name": "mycode2"

  }

]

[Gerald DeConto]

resolved.

solution was to set container_name of load_balancer node in aws_ecs_service.ecs_service  to the name of one of the containers (I used first name).

for example:  container_name = "mycode0"