https://github.com/bobtfish/tf_aws_mesos

https://github.com/bobtfish/tf_aws_mesos

Last synced: 3 months ago
JSON representation

• Host: GitHub
• URL: https://github.com/bobtfish/tf_aws_mesos
• Owner: bobtfish
• License: apache-2.0
• Created: 2015-04-05T09:46:19.000Z (over 9 years ago)
• Default Branch: master
• Last Pushed: 2015-04-11T23:12:46.000Z (over 9 years ago)
• Last Synced: 2024-07-20T03:19:16.108Z (4 months ago)
• Language: Makefile
• Size: 570 KB
• Stars: 12
• Watchers: 2
• Forks: 2
• Open Issues: 0
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

        
# Terraform AWS Mesos

## Module to set up a Mesos cluster on AWS using Terraform

See my blog post for more detailed information / walkthrough: http://bobtfish.github.io/blog/2015/04/06/diy-scalable-paas-with-terraform/

## Example Use

Create a file `mesos.tf` containing something like this:

    module "mesos" {

        source                      = "github.com/bobtfish/tf_aws_mesos"

        slaves                      = "5"

        region = "${var.region}"

        admin_iprange = "${var.admin_iprange}"

        admin_key_name = "${aws_key_pair.admin.key_name}"

        private_subnet_ids = "${module.vpc.primary-az-ephemeralsubnet}"

        public_subnet_ids = "${module.vpc.primary-az-frontsubnet}"

        domain = "${var.domain}"

        vpc_id = "${module.vpc.id}"

        vpc_iprange = "${module.vpc.iprange}"

        discovery_instance_profile = "describe-instances"

    }

See the `variables.tf` file for the available variables and their defaults

## Outputs

  * mesos_elb_dns_name - The DNS name of the externally facing ELB

  * mesos_elb_id - The AWS id of the externally facing ELB

  * marathon_api - The full URI for the marathon web interface / API

  * mesos_api - The full URI for the mesos web interface

  * domain - The domain under which applications can be requested

## Visit the web interfaces

When the cluster is set up, you need to [retrieve the NS server records for your _domain_ from route53, and add delegations

to this new subdomain from the domain you own (wherever that is managed):

![Route53 console](https://raw.githubusercontent.com/bobtfish/terraform-example-mesos-cluster/master/route53.png)

After which, you should be able to visit the admin interfaces at:

  * mesos.admin.yoursubdomain

  * marathon.admin.yoursubdomain

You can launch marathon apps using the API (or the web interface) as per [this guide](https://www.digitalocean.com/community/tutorials/how-to-configure-a-production-ready-mesosphere-cluster-on-ubuntu-14-04), and each app's PORT will be made available via HTTP at the name of the app. For example the marathon app named _/www_ becomes _www.yoursubdomain_

## To do

  * Currently no way to retrieve the address of the DNS servers for the Route53 zone through Terraform. Use the Console to retrieve the addresses. 

  * Cannot reach the log files of the Mesos slave nodes from the web interface

  * Whilst the cluster we build is redundant, currently all the machines are allocated in a single availability zone.

  * The machines have no configuration management (no puppet/chef), which means that making any changes to them (or getting any security updates) involves rebuilding the instances.

## Credits

This module is based on the code from ContainerSolutions, and their excellent GCE howto: