I have a pair of redundant Gateways (version 7.7.0) configured to use VOIP for their alarm notification. This past weekend we had an issue that caused the system to fail over to the secondary gateway. When this happened all of the alarm callouts stopped.
There was an error message that pointed me to the Local Bind Address and Public Bind Address settings to check for the problem and that is when I discovered that these settings are replicated between gateways, but they are gateway specific. My understanding is that there is no way for the secondary gateway to bind to the same IP address as the primary.
Is there a way to set these that will allow the call fail over to work correctly? How am I supposed to use these settings in a redundant system?
I may be wrong, because I usually am, but I thought I remember that the Voice Notification module was one of the modules that only worked on the master gateway. I remember reading that when 7.6 initially came out but I can’t find that document anymore.
I don’t remember reading anything in the change logs that changed that.
I would love to see some definitive confirmation or denial on this from someone at IA or in some documentation (I couldn’t find anything). As it stands this is an issue. If it is not a redundant module then the module description needs to say so and I need to be able to have independent control over the settings. Also, this raises a lot of questions about how the module functions if I have it installed in two machines that are set up as a redundant pair.
The voice module can and should work in a redundant pair, except for the issue with setting the bind address that you’ve noticed. This issue was recognized by another customer a few weeks ago, and we’ve come up with a work around, which I can get to you if you want (otherwise we’ll get it in for 7.7.3, I believe).
We are looking at a more comprehensive way for handling this type of issue. We have a few places where bind addresses are specified, and all of these (besides those on the actual redundancy settings page), would suffer from this problem.
The latest advice I’ve received (as of about 2 months ago) is to setup 2 different notification profiles. On the first you would setup the bind addresses for the primary gateway. On the second profile you would setup the bind addresses for the backup gateway. This results in each gateway having a valid configuration for their local IP address and 1 invalid one.
I hope that makes sense. It is a bit of a messy workaround.