NRS : backup NRS does not become active when primary is down

[umanitoba]

We have two networked CS1000 switches at release 7.5. When the primary signalling server is down, the backup NRS is supposed to become active but this doesn't happen. (If I gracefully disable the primary NRS then the backup does become active...but recently the primary sig server failed and the backup NRS did not work.)

Any thoughts would be appreciated!

Thanks!

 

[KCFLHRC]

So your not using Session Manager in your 7.5 system? You are still using NRS and they are deployed as Primary and Backup? What type of trunking? H323 or SIP? What else is deployed on the Primary NRS SS besides NRS?

 

[umanitoba]

No...we are not using Session Manager. We are using sig servers with one as primary and the other as backup. They are connected via SIP trunking. (Which is a problem because when the primary sig server is not operational, we lose the SIP trunking.)

 

[KCFLHRC]

I would start off looking in this NTP, NN43001-130_03.01_fundamentals_Network_Routing_Service and verify all of your settings. If this was migrated from an earlier release there are some additional settings in 7.5.

 

[umanitoba]

Thanks!

 

[golfdoctor]

Just a thought - when everything is working "normally" and you unplug the "TLan" connection of the primary NRS or stand alone NRS does the alternate NRS pick up the load ? If it does/ or does'nt then the scenario results may assist in the trouble shooting.

Curious - is the alternate NRS located at the same location as the Primary or is it at the remote site.

Again only some food for thought...


TGD

 

[umanitoba]

When I unplug the Tlan the alternate NRS does NOT take over. The alternate is located in another CS1K about 10 miles away, connected via SIP trunks over our private ethernet network.

 

[golfdoctor]

ok- just checking you left the TLan unplugged for 30-60 secs to verify ?

a few thoughts ...
double check config - mentioned previously - make sure 2md NRS is set to alternate and not failsafe per say along with proper IP settings
- was the 7.5 failover ever tested to know for sure it worked originally ? Assuming previous Nortel releases used same network connection (ie 4.9,5.0,5.5) did it actually work as advertised on previous releases.
- your remote site may be a private network but are you Layer 1 via "dark fibre" or through a managed service ( MPLS or some type of layer 2/3 service)
- Rlse 7.5 - turn on your PCAP (Primary NRS)and collect the packets for a few minutes. Both scenarios - TLAN connected and TLan disconnected on Primarary NRS - my terminolgy is probably wrong but you should see the keep alive and heart beat type stuff from the remote system.- something that the alternate NRS is verifying the Primary is alive.
- If error isolation still in doubt then if you have the equipment at the main site configure an Alternate NRS on the 2nd SS and test - you will have to change config on all SS to reflect the change. This setup should take the other variables out of play.

Only some food for thought...

TGD

 

[umanitoba]

Thanks TGD ! Yes, I left the Tlan and Elan unplugged for five minutes. I will check the settings as per your suggestion. (I've also opened a ticket with Avaya...they are investigating.)

 

[KCFLHRC]

If I remember right there was a change in the 7.5 NRS. I believe it used to be that the IP addresses for the Primary and Secondary NRS used to get reversed in the backup NRS Server. And now in 7.5 they don't get reversed, they look the same in the Primary as well as the secondary. That's what I remember anyway on a migration we did on a 6.0 to 7.5. Been a while but there was something a little different about it.

 

[KCFLHRC]

Did both sites get upgraded to 7.5, they have to be on the same release.

 

[umanitoba]

KCFLHRC: Yes, both switches are at 7.5. Thanks for your point about the IP addresses...I will look into it.

 

[KCFLHRC]

Another question, what release was this upgraded from? Was it a Linux based NRS or a VxWorks based NRS. What was the procedure followed during the upgrade? Did they backup and restore the TAR file?

 

[umanitoba]

Good questions...I will have to check with my vendor.