3 Replies Latest reply on Jan 3, 2013 11:14 AM by b124jr

    Pseudowire Counting Errors?

    b124jr New Member

      We are serving a single customer who has multiple locations with PRI circuits being delivered over 408 devices. There is a Central Office 408 and each site has a 408 that has L2 connectivity back to the Central Office device for connectivity between the pseudowire interfaces.

      Site A and Site B have been operating clean for the last year. We have just recently turned up Site C and have had an issue with the PRI circuit staying up. The circuit will only come up for about 1 minute and 30 seconds if I bounce the VLAN interface.

      I have attached a drawing and the configs from each device. The only evidence on the Site C 408 I can see is pseudowire 5 counting missing packets and jitter buffer errors. I have also attached a log of the counters I am seeing on the Site C device.

      I can ping continuously across the VLAN 705 interfaces with sub 10ms response times. Is there any clear evidence of what is going on across Site C’s circuit?

        • Re: Pseudowire Counting Errors?
          b124jr New Member

          Here is drawing

          • Re: Pseudowire Counting Errors?
            adtn99 Employee

            b124jr,

             

            I suggest performing a "show timing" on site C's MX408e to secure the PDV which will provide you with the average packet delay across the ethernet network.  This value is used to set the Jitter Buffer Depth for the Pseudowire in question.  If the PDV is over 100000uS please peform a "clear counters" wait 30 minutes and perform another "show timing" to secure the current PDV reading.  For example if the current PDV = 23000uS  I'd suggest setting the Jitter Buffer Depth of the Pseudowire to 28000uS clearing the counters again and let it run for 24 hours.  This should eliminate/reduce the Pseudowire errors caused by delay.

              • Re: Pseudowire Counting Errors?
                b124jr New Member

                After issuing the show timing command, the PDV value was way beyond 100000. There was a media convertor at Site C in front of the 408e. We changed to direct fiber to the 408e device and now the PDV values are within range that are handled by 408e.

                Thanks for the advice.