Skip to content

Latest commit

 

History

History

ha-cluster-existing-vpc

README.md

Deployment of SNS EVA HA cluster on Outscale in existing VPC

Introduction

This Terraform script deploys a SNS EVA high availability cluster on Outscale in a single availability zone in an existing VPC.

Schematic

Requirements

  • Terraform
  • Outscale Terraform Provider
  • Outscale VPC with 3 subnets, an Internet Gateway, a default route table with a default route
  • SNS version 4.7.1 or greater

configuration

  • Edit the variable template terraform.tfvars.example

$ cp terraform.tfvars.example terraform.tfvars

  • Add your Outscale access key/secret and key pair name to terraform.tfvars
  • Configure the three subnets:
    • Public : Internet facing subnet
    • Private: subnet with the assets to protect
    • Sync: subnet dedicated to appliances synchronisation
  • Configure the ip addresses according to the subnet networks

Deployment

$ terraform apply
$ terraform apply

Note: The two apply commands are necessary for the source/dest check parameter.

Terraform outputs the main EIP of the cluster and the temporary EIP for the second EVA VM.

Cluster creation

  • Connect to both VM web gui to install licences/init kit.
  • Create the HA cluster on the first EVA VM
 > CONFIG HA CREATE ifname=sync password=<secretpassword> Unicast=1 SynchronizeMacAddress=0
 100 code=00101e00 msg="Success"
 > CONFIG HA activate
 100 code=00a00100 msg="Ok"

  • For firmware version lesser than 4.7.1, the Unicast and SynchronizeMacAddress options don't exist. It is necessary to delete the forced mac address from the configuration. Via SSH access edit /usr/Firewall/ConfigFiles/network file and delete or comment the MacAddress tokens of the ethernet0 and ethernet2 sections.

  • Connect to the second EVA VM via the web gui and join the cluster on EVA2

> CONFIG HA JOIN ip=10.0.2.11 password=<secretpassword>
100 code=00101e00 msg="Success"
> CONFIG HA Activate
103 code=00102700 msg="rebooting..."
Leaving client...

  • Delete the temporary EIP for EVA2 which is no longer used

  • Connect to the cluster IP via https, the high availability monitoring will update the cluster status after the second VM reboot.

Clean-up

To remove all deployed ressources:

 $ terraform destroy