ISDN PRI - E1 problem

Answered Question
Feb 19th, 2007

Hi all,

I have VWIC2-2MFT-T1/E1 in my voice gateway (2821). These are configuration on provider's side :

- isdn switch type Euro

- framing no-crc4

- linecode hdb3

- network side

- d-channel 16

- timeslots 1-31

Here is my configuration :

network-clock-participate slot 1

network-clock-select 1 e1 1/0

isdn switch-type primary-net5

controller e1 1/0

framing no-crc4

linecode hdb3

pri-group timeslots 1-31

no shut

int s1/0:15

isdn switch-type primary-net5

isdn incoming-voice voice

no shut

So far, the link is always up and down. I see that layer 2 is sometimes up before get RAI or LOF detected message and down. Cisco documentation said to try loopback test and change another port. I have tried both of them, but still can't find solution. The provider said that they use Lucent on their device. does anyone experience with Lucent and Cisco compatibility ? See attached file on debug events, debug isdn q921, and debug q931.

Regards,

suwa

I have this problem too.
0 votes
Correct Answer by Paolo Bevilacqua about 9 years 7 months ago

Hi,

so we can observe that for neither the 5300 or the 2801 have a clean circuit. See how the 5300 keeps retrasmitting UAf and it seems the switch is not receiving it as it sends SABME again and again.

Now for some reason (different harware, etc) seems like the 5300 is less sentitive to errors and can marginally work.

I still think that either the circuit, or the port on the telco switch is faulty, and if it was me, I would not sign acceptance of service.

I had a similar case and it tooks days of argumenting with telco until the issue finally got solved. It was a faulty port on the switch but they didn't want to admit that. I'm not saying the problem is the same but the attitude is, they were unable to produce a circuit analyzer but here you are proving them with two different router that is not working.

Good luck!

  • 1
  • 2
  • 3
  • 4
  • 5
Overall Rating: 4 (1 ratings)
Loading.
Paolo Bevilacqua Mon, 02/19/2007 - 04:08

Hi,

Is this a circuit to telco, or a local connection?

How the circuit is terminated ? Is there a box (sometime erroneously called modem), and which cabling are you using to the termination ?

However, if a circuit, the onus of testing it, including loopback etc, is on the provider.

0600648902 Tue, 02/20/2007 - 01:04

Hi,

It's a telco provider. provider has tested loopback and they said that everything runs ok. There are 2 cable (tx and rx) with 2 lines each (so called tip and ring). It is terminated through rj45.

Recently, the provider said that it might be something wrong about isdn timer parameter. Any idea ?

Thanks.

Paolo Bevilacqua Tue, 02/20/2007 - 03:43

Hi,

I have looked at the trace. It is evident that the circuit is so severely errored that no reliable link can be established, hence L2 flaps continuously. In particular, it seems like the switch is either loosing or receiving errors from the router, as it continuously tries to reinitialize L2.

I only have few tips for you:

- It seems strange to me that the clock is show as "derived from backplanes" in show controller. Please configure clock source line

- Check cabling. Make sure you do not use ethernet patch cables, but twisted pair with the correct pinout. Ethernet is twisted differently and should not be used, espcially for long runs.

- upgrade IOS to latest 12.4 or 12.4T (just as a caution).

- All failing, put the burden of proof on the telco. Ask them to bring a circuit analyzer to be inserted between router and circuit termination. Observe for errors in both directions. If you see that the router sends no errors (as 99.99% of the cases), then the circuit is faulty.

I don't think that changing L2 timeous would help. Unless you have a clean show controller, this link will never works.

0600648902 Wed, 02/21/2007 - 01:54

Hi,

1> I have configured clock source line, I'll check it later whether the "derived from backplanes" is still available or not.

2> I don't know which cable provider is using, but provider is using the same cable to other isdn pri clients.

3> I'll try it later

4> I'll try to ask them using circuit analyzer.

Just for info: Provider ever tried to change the framing into CRC4. The time they changed the framing, lots of alarms reported to their lucent device.

I'm sure packets/frames are lost on their way to provider's device (from the "turning on crc4 framing" accident). what do you think ?

Thanks

Paolo Bevilacqua Wed, 02/21/2007 - 02:16

For sure both sides should agree on CRC4 / no-CRC4, but I think you have tried this already.

Now, because "clear counters" does not clear counter at conotrller level, your only option is to reboot the router to start again monitoring. But the fact alone that the circuit intermittently reports LOF/RAI is an indication of a fault.

Remember that in the best tradition of telco internationally, they will not admit there is a problem until is demonstrated to them the hard way.

0600648902 Thu, 02/22/2007 - 01:53

Hi,

Of course, I've set crc4 on both side. I monitor the controller counters, I found that the linecode violations and pathcode violations increased slightly. I searched cisco.com and it said that it might be caused by physical error. Is it possible to try the connections using isdn phone? if isdn phone works, the error is on my side, isnt it?

Thanks

Paolo Bevilacqua Fri, 02/23/2007 - 15:49

Hi,

I don't think there are ISDN phones made for PRI, for sure there are for BRI.

Rest reassured that when the clock is set correctly, and the circuit are clean, all error counters stay to zero and ISDN L2 is stable. Cisco has probably hundreds of thousand of these installations and in our experience, when the basics of a circui don't work, is not router fault.

0600648902 Tue, 02/27/2007 - 19:52

hi,

update info: e1 is up when I use AS5300. L2 status is multiple_frame_established and no error on interface counters. But something I want to ask, is there something wrong when the debug isdn q921 shows that tx sent UAf and rx received SABME all the time? From cisco site, it said that there is something wrong with L2 negotiation if SABME showed up. But L2 was stably up. Any idea?

I'm trying to download the latest IOS,named c2800nm-spservicesk9. Let's see ..

0600648902 Tue, 02/27/2007 - 22:00

Hi,

upgraded IOS do not make things better. I want to ask one more thing, I heard that AS5300 was smart device that was able to detect any parameters from the telco side. Is it true? if yes, please tell me how to get the parameters.. thanks.

Paolo Bevilacqua Wed, 02/28/2007 - 06:43

Hi,

well, the fact the 5300 is working is good, and indicates that something must be wrong with the other router. But you said that you have checked and rechecked impedance, clocks.. etc to no avail.

I'd say the interface that was not working should be tested in loopback or with a circuit analyzer before being declared faulty.

Re: SABME it is perfectly legal and possible that SABME/UAf are sent and received indipendently and at the same time. If you tell us where did you read that's not the case, one can try to have that corrected.

Re: 5300 Vs other router: The only way it is smarter, is because it can terminate data/voice/modem/fax using a single platform (universal port). However, the ISDN software and parameters are perfectly identical to the other platforms as you will see using it a little.

0600648902 Wed, 02/28/2007 - 22:11

Hi,

Loopback test has been done with this procedure:

http://www.cisco.com/en/US/tech/tk713/tk628/technologies_tech_note09186a008010059a.shtml

everything is doing well.

there is no circuit analyzer here.

When using 5300, there is a little error,not much, when I let the E1 up for several minutes. I attached some files for comparison between 5300 and 2821.

Thanks.

Attachment: 
Correct Answer
Paolo Bevilacqua Thu, 03/01/2007 - 10:22

Hi,

so we can observe that for neither the 5300 or the 2801 have a clean circuit. See how the 5300 keeps retrasmitting UAf and it seems the switch is not receiving it as it sends SABME again and again.

Now for some reason (different harware, etc) seems like the 5300 is less sentitive to errors and can marginally work.

I still think that either the circuit, or the port on the telco switch is faulty, and if it was me, I would not sign acceptance of service.

I had a similar case and it tooks days of argumenting with telco until the issue finally got solved. It was a faulty port on the switch but they didn't want to admit that. I'm not saying the problem is the same but the attitude is, they were unable to produce a circuit analyzer but here you are proving them with two different router that is not working.

Good luck!

0600648902 Thu, 03/08/2007 - 21:46

Hi,

I request the provider to re-do BER Test. The provider loop the link over my router side. After doing BER-Test, I don't know what makes the E1 stably up, no detected error, no changed configuration, and eveything is still the same with the time it was flapping. I believe that there was something wrong at the provider side and got it fixed during the BER Test.

Anyway, Thanks a lot for your kind guidance, p.bevilacqua! Nice to meet you here.

Regards

Paolo Bevilacqua Fri, 03/09/2007 - 08:17

Thanks to you for giving the news of an happy end.

And for the confirmation that certain attitudes never change, nothing is ever wrong with telco services before escalation. And sometime, not even after :)

0600648902 Mon, 03/26/2007 - 22:29

An unhappy event showed up a few days after being stable. Yes! the link went flapping again. After searching it for weeks, finally I got (from my senior) a TAC Case looked alike with my case. It said that the module had to be configured as normal mode. The default of my card was wetting current mode. Since turning to normal mode, issue has been solved. Just check it:

http://www.ciscotaccc.com/voice/showcase?case=K84133319

Hope this post able to help others with the same case.

Regards.

Actions

This Discussion