Multi site IPO having 3-digit dialing issues with Comcast MPLS service

[1967LS6]

Im hoping one of you really sharp fellas has some ideas for me. After a couple months we are at our end on this one; So we have a customer with 6 sites with IP office, using the Comcast's version of an MPLS network.(i forget the exact term they call it) So they each have a control unit running 8.1.xx, with about 10-15 phones at each site. Each site has a Sonicwall, all identical units as far as I know. Whats happening is that at random time frames, ranging from 1 day to 3 days between occurances, they lose the ability to 3-digit intercom dial from site to site. The IPO shows no errors, no alarms for the SCN, NOTHING. But they get a busy signal when they try to dial site to site using the "network". They also of course lose access to their mail server (VM pro) which is at what we will call the "Main" site. So basically they lose network connections. However, we see no errors as i said, AND we can "ping" from one Control unit to another, in any direction, any sites, and everyone "sees" each other, but no phone to phone intercom ability. But if we just restart the control unit at the main site, it starts working again just fine, for a random amount of time. last week it went 4 days just fine. then once it happened twice in one day. The IT tech for the client has spent countless hours wth Sonicwall support, Comcast support, etc. checking settings, etc. but no luck. We even resorted to switching out the two combo cards in that particular site to rule out a "glitch" in the VCM funstions of the module. We are going to switch out the entire control unit tomorrow. All of us including our programmer does not feel its an issue with the IPO or losin its network connections, because like I said, we can ping back and forth so the TCP seems fine when the 3-digit diaing is not working. Also whats strange is that none of the 6 sites can communicate but only resetting the "main" site fixes everything. Everyone we talk to says 'its the firewall"...but what??? And if it was a setting, either in the firewall or the IPO, the settings do not change so it would either always work or not. Its like it hits some sort of "threshold" then freezes up. I have no other clients using this Comcast type multi-site service to compare to. One last thing: in case it matters, they have a full PRI at the "main" site, and all the other sites are programmed to dial out THROUGH this site on thr PRI. each of smaller sites only has analog lines fro inbound only. So all of this outbound dialing ability lost during this issue....HELP!!!!

 

[Gunnaro]

You say R8.1.xx, are they on different versions of the R8.1?
I would start by getting them on the same level if they aren't, and make sure your VMPro is on the corresponding release.

How is your SCN setup? Hub with single spokes to the satellites? Make it mesh, then you can communicate inter-site even if main site goes down.

I can only guess what's causing this:

- IP conflict
- Faulty HW (IPO, switch, cabling..)
- Config error



Kind regards

Gunnar
__________________________________________________________________
Hippos have bad eyesight, but considering their weight, it’s hardly their problem

 

[1967LS6]

Thanks for your input. It is set up as a "mesh" with IP routes built from every site to every other site, and so on. but when the site to site ability just fails (like I said random times, 1-3 days between), all the sites lose their ability to connect to any other site. Then rebooting the ONE site resolves the issue. The one site if the one with the PRI though. But regardless, what would the PRI have to do with trying to intercom dial from site to site??
to answer your questions:
1. yes, they are all running the same version. All the units were installed at the exact same time.
2. if it was a faulty switch or IPO, why would thr TCP still work fine, just lose calling ability? (when the problem happens, we can still ping sites and also ping te control units from the VM server, etc.
3. If it was a config error, would it not either work or not work? it would not work just "most of the time", but not all the time....
????

 

[Gunnaro]

Hardware faults on IPO could be the LAN card(s) or bad software, missing files, etc.
Strange things can happen when mixing on R8.1.

If you built the IP lines for every site-to-site and routing is ok, you should have communications internally.

It's pretty difficult to troubleshoot something like this, without seeing the full picture.
I think I would start by replacing the Main site HW, all of it if you have the spares to do it.






Kind regards

Gunnar
__________________________________________________________________
Hippos have bad eyesight, but considering their weight, it’s hardly their problem

 

[IPGuru]

General rule for trouble shooting VoIP links between sites

IT IS the data network!

A Maintenance contract is essential, not a Luxury.
Do things on the cheap & it will cost you dear

 

[Gunnaro]

...And when one has the luxury of having a complete spare IPO to put in, it's easier to claim that after you tried everything...
Then you send a bill for your services

Kind regards

Gunnar
__________________________________________________________________
Hippos have bad eyesight, but considering their weight, it’s hardly their problem

 

[yankblan]

Paragraphs please, too!

ACSS-SME

http://interconnexionsld.ca/

 

[yankblan]

Do you have combo cards on all of them?

Take them out and put in VCMs, done this on a couple of installs and it resolved either/or SIP and SCN problems, after trying everything else (SD, main unit, cards, ect)

ACSS-SME

http://interconnexionsld.ca/

 

[1967LS6]

Thanks for the input guys. There are two combo cards in the unit. We have already switched out one, we are switching out the other and the control unit itself today. other than that, other than that there is only a PRI and Dig 8 in the box...