Freepbx - 2 systems - howto use 2nd system as automatic failover?

i have my main system Freepbx15 running in a VM on my server
it has 2 trunks, i for business, 1 for personal

i have raspbx15 running on a Pi
i have added the IP of my Pi to my 2-line voip phone, and my ATA
i see that these extensions are now registered with both systems

at the moment i have the 2 trunks disabled in my Pi
for testing i disabled the trunk in my main system before enabling on the Pi

for automatic failover to work i need the trunks connected from both systems
will this be a problem with my providers ( & freephoneline)?


You can’t have automatic failover. as the provider will only send the call once.
If both systems are registered, you are basically on random chance which one it will send to. Typically it is last to register.

What is the real goal here? Do you not trust your VM infrastructure? I mean that seriously. A properly setup VM infrastructure means you can be back up and running from a backup in minutes.

Obviously, your needs for HA are not real because you are attempting to do some fake HA with a Raspberry Pi.

many vsp provide for a failover destination for your dids if the primary one fails. analagous to your proposed server arrangement.

1 Like

sometimes when i apply updates to my server, it is down for over an hour (Microsoft, what can i say)
sometimes i am updating hardware and have to take it down for more than a few minutes
in these cases i would like my voip system to be up on the Pi
my ATA has ‘primary preferred’ so i am hoping it uses that as long as it is up
my Grandstream 1625 voip phone just has primary/secondary, but says i am registered to my main server


jared ive evolved my view of HA for voice

i dont think of it as insulating the hardware anymore but insulating the service

with that in mind a valid means to achieving voice availability is one that uses redundant registrations on the endpoints , preferably IP and that also utilizes an ITSP’s ability to have alternate routes should the primary one fail; if you abstract the CDR from each PBX to a remote sql instance and utilize voicemail to email you can construct a extremely fault tolerant voice solution and if one wants to take it a step further and insulate the services accessed via fqdn something akin to route53 and health checkers are a great solution

anyway not sure if you would consider that fake HA but the above scenario works pretty well and allows for site level fault tolerance

ive got my replication scheduled at every 15 minutes with the new process in freepbx 15 , check it out someday … we need for feedback


+1 @dolesec

Adding corosync to asterisk is relatively easy and allows for each server to know the state of the other server and share any ‘state’ prior to an enforced ‘swap’ , personally i use inotifywait to intelligently syncronize /var/spool/asterisk/*

This is the best quote ever when it comes to HA:

High Availability is not something that you buy, it is something that you do. - John Nicholson

This topic was automatically closed 7 days after the last reply. New replies are no longer allowed.