Dual MCD, Dual MBG, incoming calls only work on one MCD

[handwritten]

We have two MCD installs: one MXe and one VMCD, both running 7.0 PR1.  We use SIP trunking with two MBG units running V8.0.27.0.  Both MBG units are configured for both MCDs.  During a recent power disruption that lasted 10 minutes*, phones failed over from the MXe to the VMCD as expected, but incoming calls received dead air.  Outgoing calls were fine.  How should I go about troubleshooting this?


*By the way, I expected that having two power supplies would maintain service during the replacement of one power supply (this is mentioned in the documentation).  However, when power is lost to one power supply, the unit reboots when power is restored.  Has anyone else seen this?  Is this a feature of the platform, or do I have a faulty unit?

[acejavelin]

We have two MCD installs: one MXe and one VMCD, both running 7.0 PR1.  We use SIP trunking with two MBG units running V8.0.27.0.  Both MBG units are configured for both MCDs.  During a recent power disruption that lasted 10 minutes*, phones failed over from the MXe to the VMCD as expected, but incoming calls received dead air.  Outgoing calls were fine.  How should I go about troubleshooting this?


*By the way, I expected that having two power supplies would maintain service during the replacement of one power supply (this is mentioned in the documentation).  However, when power is lost to one power supply, the unit reboots when power is restored.  Has anyone else seen this?  Is this a feature of the platform, or do I have a faulty unit?

Do you have redundant inbound routes for inbound SIP calls to send them to both PBX's from both MBGs?

As far as the power supply things, I have changed a single PS unit in a dual system more than once, I don't recall a reboot ever occurring.

[handwritten]

Redundant inbound routes.... hmmm... that sounds like an excellent thing to have.  I can't find any call routing setup on the MBG units.  Would it be configured on the MBGs or on the MCDs?

[acejavelin]

Redundant inbound routes.... hmmm... that sounds like an excellent thing to have.  I can't find any call routing setup on the MBG units.  Would it be configured on the MBGs or on the MCDs?

Inbound routing to each controller is handled by the MBG if it is your SIP SBC (most likely), go into SIP Trunk edit and go to "Rules" at the bottom and set the Primary and Secondary controller as your defined ICP's. Of course, you would need duplicate DID routing (or make it a shared form) in each controller's DID Routing form as well.

Of course, the alternative would be to have your carrier setup the DID's in a fail-over, if the call can't complete to MBG 1, they send it to MBG 2, but then all routing must be duplicated on both MBG's anyway.

Or for true redundancy, have both.

[handwritten]

I have a rule that matches 'Request URI' as your screenshot shows, but the rule field just has a '*' (indicating all rules?).  Both MCD systems are set properly as primary and secondary.

[handwritten]

The rule as I describe appears appropriate, as per the MBG Installation and Maintenance Guide:

"• * a routing rule consisting of a single "*" character allows any pattern to match. "

[acejavelin]

The rule as I describe appears appropriate, as per the MBG Installation and Maintenance Guide:

"• * a routing rule consisting of a single "*" character allows any pattern to match. "

That's fine... our test system as shown is different as we have multiple different systems (MiVoice Business and MiVoice Office 250) sharing the same SIP trunk and a single MBG, but the Primary and Secondary would be defined the same way. Just define both of your PBX's in ICP's and change the drop-down menus.

If it is failing, what do the traces look like? Is ICP-1 rejecting/failing the call so the MBG can send it to ICP-2? Is ICP-2 accepting the call?

Best bet would be to cause a failover condition outside of normal hours and do some packet captures in the MBG and see where the call is failing... I can tell you if all the IP routing and redundant DID programming is in place it works, even across different platforms.

[handwritten]

That's great - thanks so much for your help.  I will do as you suggest and post back if I learn anything that would be of use to others.