MGC Daughter board issues!!!

[MikeDThePhoneGuy]

I'm installing an NTDW64 daughter board onto the NTDW60 MGC and for some reason the board will not work, at all, for the first hour...then suddenly it's up and running...All IP's and programming in the loads is/are correct...Any thoughts...I'm stumped. Thanks!

 

[MikeDThePhoneGuy]

Anyone...anyone...Beuler...Beuler...

Seriously I'm stuck for some kind of help.

 

[sl1input]

Is it on Rls 6.0 or above?

 

[MikeDThePhoneGuy]

No...The Switch is Rls 5.5, and the MGC is Rls 2

 

[Tman45]

So what is your real problem - if you say it wasn't working for the first hour and then magically working? We have a 5.0 system with MGC release versions 1-3 and different vintages of daughterboards. I would just watch that one you had trouble with to see if the DSPs go offline and if you continue to have problems with it then let us know we can try to help.

 

[MikeDThePhoneGuy]

Sorry for the lack of details so let me explain.

We are in a Lab environment and can test a number of DW62's and DW64's. For some reason none of the DB's will work when first installed not even the ones we use full time in our 1000E. Usually the problem is no voice path from TDM to IP but all VGW's are idle.
After 15/20/30 minutes(no set time) the DB's will be fully operational with voice path both ways. They pass hardware tests in EM and the boot on the MGC is fine it's just that voice path issue for a certain amount of time when the DB is replaced.

 

[bhassell2000]

is it possibly your MGC card in your lab switch?

Thanks,

Buddy

Linked in Profile

http://www.linkedin.com/in/buddyhassell

 

[MikeDThePhoneGuy]

How so?
After the DB's do come on line they will stay up and work fine until the next time I have to test a DB.
Needless to say the MGC has been put in and out quite often but it boots fine.
Do you know of any issues with firmware/loadware in regards to the DB's and MGC?

 

[bhassell2000]

have you put the MGC from your production PBX into you lab and duplicated the issue? IS it always the same MGC card?

Thanks,

Buddy

Linked in Profile

http://www.linkedin.com/in/buddyhassell

 

[CollegeFootball]

Always the same MGC.

 

[Romar53]

I ran across this problem when doing upgrades to 6.0.The first time a call is placed across the DB you get no way speech path. To get around this I disabled the 32 port DB, made a call across the 96 port DB. Keeping the call up and wait about 30 seconds the call will establish. It then will process calls. Make sure you hit each DB. This happened on about 12 upgrades. I don't have a reason why, it only happens on the first call.

 

[CollegeFootball]

My problem is a bit different.

Same MGC card for all testing.
Will install DW62's &DW64's when required to test.
After completion of MGC boot all VGW's are idle.
On first test there is no voicepath from TDM to IP.
After an amount of time 10/20/30 minutes I will test again and unit is fine.

Not sure why it delays. There are no errors output during this timeframe.

 

[biv343]

My guess is a network issue. Most layer 3 switches keep a cache of IP address to MAC addresses in what is called the arp cache table or forwarding database. You wont get any speech path until the data gear and other equipment learns the MAC of the new daughterboard, which could take 5 minutes or more depending on how the gear is configured.. You could try clearing the arp cache on the data gear next time this happens and see if that resolves the problem.

 

[Tman45]

I would agree with biv343 as I have seen this same issue with IP phones and Cisco network gear (even Cisco and Nortel phones).

Good luck on your fix.

 

[fritzthecat]

Hello,
I would say it is definitely an arp cache issue, very easy to verify, just clear the arp cache at the switch side before you insert the exchanged board, we had that very often and most of the time cisco switches are having that behavior

 

[dgjlee]

Hello,

We are having a very similar issue that is down to the DSP resources dropping our of the arp table on our switches. We are seeing this on Nortel 5520 and 8300 switches. Did you ever resolve the issue?