Does anyone know if it’s possible to setup a warm spare freepbx instance that would be tied to an HA pair in another location? The idea would be to have a disaster recovery setup in case the HA pair site goes down, the warm spare would have the current configuration and could be made primary. If possible, would these instructions be correct: https://wiki.freepbx.org/display/FPG/Warm+Spare+Setup
If that’s not possible, are there any suggestions for putting the above scenario into place?
Thanks
I don’t see why not.
The Warm Spare will be connecting to one of the HA machines and pull a backup.
With that being said, Sangoma is working a on a WAN HA solution.
Yeah, I’m a little interested to see how this is done because I don’t think it’s going to be as magical as people think. The whole thing with HA is that there is a floating IP address, that’s all good and well when you’re on the same subnet and LAN but when it’s over the Internet that’s when it gets interesting.
I put a PBX in the East DC and a PBX in the West DC. How do they link? How does that Floating IP get from East PBX to West PBX? How are they monitoring themselves? Those are just a small example of other questions/items that need addressing.
Now I’m not asking because I don’t know how it could be done, I do. I’m just trying to make people aware of some things and not continue to think it will be a automagical solution. I could see people trying to set this up and discover they may need certain things like an ASN or BGP routing depening on their setup.
There are already ways to do this without the need of an HA module. I am still interested to see how this solution will be handled in regards to the HA module.
It won’t be perfect, but it will be good enough. You’re not trying to serve a 911 PSAP, an air traffic control center or a nuclear plant control room.
I believe it’s a given that calls in progress at the time of a failure will be dropped. Geographic separation won’t make that worse.
It’s also understood that if the primary doesn’t crash but gets ‘sick’ in certain ways, the system won’t be usable until a human forces the failover. Better software will reduce the window, but it will never be eliminated. IMO, separation doesn’t make that significantly harder.
Using a DNS SRV record, even a static one, the backup system would be fully online within about two minutes after a hard failure of the primary. With some careful call routing planning, calls entering the backup during this interval can be queued and delivered to agents once they have registered. Customers would see nothing worse than a longer wait time.
So to confirm the original question, there should be no issues connecting the warm spare in the off-site location to the production HA cluster using the instructions in that link?
Thanks
In theory, no.
Ok, thanks everyone!
This topic was automatically closed 7 days after the last reply. New replies are no longer allowed.