tuxomatic-ovhipfailover

Module to manage a network interface script and manage routing of IP failover on OVH servers. Meant to run on RedHat based systems. Management of OVH IP failover (with routing accross multiple datacenters) is handled via an init script. The script is compatible with cluster resource managers such as rgmanager or Pacemaker to offer a complete HA stack by nesting it under a VIP resource.


Keywords
redhat, ha, network, ovh, failover
License
GPL-3.0
Install
puppet module install tuxomatic-ovhipfailover --version 0.3.2

Documentation

OVH IP failover

This Puppet module is meant to ease the implementation of multiple network interfaces on a server and leverage special network solutions offered by the hosting company OVH.

Network interface via RedHat type network script

  • General network interface, private/public.
  • Virtual interfaces, included the ones offered by OVH as part of their vRack solution.

Manage a failover IP from OVH via an init script

  • Wraps underlying calls to the OVH API to control the routing destination of the IP (works between different data centers).
  • Can control virtual interface state regardless of subnet which isn't possible with OCF scripts such as RedHat's rgmanager.

Example

You can manage in a single definition a virtual interface with failover capability provided by OVH. Apply the definition on serverB by just changing the destination_fqdn to "serverA" and device name if differently mapped.

 class { 'ovhipfailover':
       ipaddress => "x.x.x.x",
       device => "ethX:X",
       destination_fqdn => "serverB.domain.com",
       application_key => "",
       application_secret => "",
       consumer_key => "",
 }

To obtain a consumer key to use on the OVH API after creating an application key and secret, use the embedded Python script. Calling the script will return a consumer key and a link to the validation url, you should choose an "illimited" time validity for your consumer key.

./ovh_ip_failover.py POST /auth/credential <APPLICATION_KEY> <APPLICATION_SECRET>

By using the failover init script in a cluster resource manager such as rgmanager or Pacemaker, a true HA solution can be implemented across multiple OVH datacenters. No extra parameter is needed to call the script as a standard init script resource.