Autoscaling in the cloud and Internal IP

[tmediassembly]

Hello,

I recently virtualized a FusionPBX server and noticed that the internal IP had, of course, morphed with the 'VIRTUALIZED IMAGE', into a new IP. This caused some INTERNAL errors on the VIRTUALIZED IMAGE, so I decided not to make any new servers like that (via virtualized images or snapshots).

So, can you tell me how and if FusionPBX would work with autoscaling in AWS? I am intersted in seeing how others have dealt with it. If there is a good post to read or issue, please let me know as I have not seen any related to AWS. I am specifically talking about the load balancers, autoscale groups, and INTERNAL ips in FusionPBX.

Thanks!
G

 

[mat1010]

Hi, the usual process with autoscaling in AWS is, that you automate the config generation which is specific to the added instance within the startup process by either using scripts through userdata, or builtin mechanisms within your application. to retrieve the required information. As an example, you could fill the vars.xml by reqúesting AWS' metadata service to retrieve a public ip:

<X-PRE-PROCESS cmd="exec-set" data="bind_server_ip=curl -s http://instance-data/latest/meta-data/public-ipv4"/>

Also documented here. Afaik there's no documentation how to build a resilient fpbx setup with autscaling included on AWS since this does require a lot more than just scaling the fpbxs and address them through a L4 loadbalancer. You would need some kind of SIP proxy (kamailio, opensips) or dialplan logic within FreeSWITCH that takes care to ensure that calls, call pickups, conferences and so on can find each other.