> On Fri, 2009-06-12 at 10:03 -0400, Scott Lawrence wrote:

> What isn't supported or tested is having two members of an HA cluster
> behind two different NATs.  If what you have is two different systems
> (different SIP domain names) configured to talk to each other via
> site-to-site links, being behind two different NATs can be made to
> work
> (with the usual NAT hassles, of course).
> 
> On Tue, 2009-06-16 at 11:39 -0400, Matt Keys wrote:
> 
> Scott,
> 
> Can you clarify this? Lets say server A is behind a NAT but in the
> DMZ. A new branch office will have server B also behind a NAT. You're
> saying they must be in different SIP domains? HA isn't available
> without VPN or similar connectivity?

That is not a configuration that we expect to work (and at this time
would not consider it a bug if it didn't).  (Note that "in the DMZ" is
not a statement that has a well-defined meaning, which just scratches
the surface of the ambiguities in this problem space)

There are a bunch of interactions between systems in an HA cluster.
Making them all work through NATs would be a lot of work.  More
importantly, setting up and maintaining a regression test process that
ensured that remained true as we modify and extend those interactions
would be an even bigger job.

By contrast, setting up a VPN between those same two sites so that the
systems did not have to worry about it is something that can be done
using well supported off the shelf technology.


_______________________________________________
sipx-users mailing list [email protected]
List Archive: http://list.sipfoundry.org/archive/sipx-users
Unsubscribe: http://list.sipfoundry.org/mailman/listinfo/sipx-users
sipXecs IP PBX -- http://www.sipfoundry.org/

Reply via email to