ketzacoatl

Install ruby-build

dep

sudo apt-get install -y libreadline-dev

src

ᐅ git clone https://github.com/rbenv/ruby-build.git

ketzacoatl

locals {
  # we have to fixup the master username before we use it, the AWS API will error out otherwise
  # RDS expects the username to be without hyphens and 16 characters at max. 'rdsadmin' is also
  # a forbidden username, though we don't validate that here.
  db_master_user_unsanitized = "${var.app_db_master_user}"

  # truncate to 16 characters, but deal with the fact that substr() will error out if you ask
  # for more characters than are in the string
  default_master_user_max_length = "16"

ketzacoatl

# Security Group for ALB
resource "aws_security_group" "atlassian-alb" {
    name = "${var.name}-load-balancer"
    description = "allow HTTPS to ${var.name} Load Balancer (ALB)"
    vpc_id = "${module.vpc.vpc_id}"
    ingress {
        from_port = "443"
        to_port = "443"
        protocol = "tcp"
        cidr_blocks = ["0.0.0.0/0"]

ketzacoatl

# Grafana deployment
#
# TODO: Auto-import data sources and dashboards. See for example:
# https://raw.githubusercontent.com/giantswarm/kubernetes-prometheus/master/manifests-all.yaml
#
# After bringing up:
# * add datasource: prometheus http://prometheus
# * import dashboard: https://grafana.net/dashboards/315
#
apiVersion: extensions/v1beta1

ketzacoatl

apiVersion: v1
kind: ConfigMap
metadata:
  name: prometheus
  namespace: ops
data:
  prometheus.yml: |
    global:
      scrape_interval: 30s
      scrape_timeout: 30s

ketzacoatl

This is the common component:

# Create a single load balancer for all Atlassian services
resource "aws_alb" "atlassian" {
  name            = "${var.name}"
  internal        = false
  idle_timeout    = "300"
  security_groups = [ 
 "${aws_security_group.atlassian-alb.id}",

ketzacoatl

Keybase proof

I hereby claim:

• I am ketzacoatl on github.
• I am ketzacoatl (https://keybase.io/ketzacoatl) on keybase.
• I have a public key ASABHBuq9eKf_Tg1I5nNIxbqYyoqLkQixEdaZ1v_cOSNcAo

To claim this, I am signing this object:

ketzacoatl

2016/11/19 09:49:12 [INFO] Terraform version: 0.8.0 dev 283d49f12fe203a2cd2ad23acec046f72f7842de+CHANGES
2016/11/19 09:49:12 [INFO] CLI args: []string{"/home/user/bin/terraform-v0.8.0-dev", "apply"}
2016/11/19 09:49:12 [DEBUG] Detected home directory from env var: /home/user
2016/11/19 09:49:12 [DEBUG] Detected home directory from env var: /home/user
2016/11/19 09:49:12 [DEBUG] Attempting to open CLI config file: /home/user/.terraformrc
2016/11/19 09:49:12 [DEBUG] File doesn't exist, but doesn't need to. Ignoring.
2016/11/19 09:49:12 [DEBUG] Detected home directory from env var: /home/user
2016/11/19 09:49:12 [TRACE] Preserving existing state lineage "a56754c9-aed5-4dce-92f8-921e3c1de105"
2016/11/19 09:49:12 [TRACE] Preserving existing state lineage "a56754c9-aed5-4dce-92f8-921e3c1de105"
2016/11/19 09:49:12 [TRACE] Graph after step *terraform.ConfigTransformerOld:

ketzacoatl

What..

This is a minimal reproduction of an issue I am seeing when provisioning hosts with Saltstack and Packer. More specifically, running salt-call --local saltutil.sync_all in a script uploaded by packer does not function as expected (no files/modules are sync'd).

Note that saltutil.sync_all works fine on the new host when run manually from the shell directly.

How to use..

• Download and install packer to your $PATH from https://www.packer.io/downloads.html

ketzacoatl

Useful Snippets for a few Random Ops Tasks