[c-nsp] Incorrect bandwidth

[Andy Koch]

On Tue, Mar 9, 2010 at 03:26,  <nasir.shaikh at bt.com> wrote:
> Hi,
> I have an 2621XM running c2600-ik9s-mz.123-22a.bin and I noticed
> something strange.
> Reports were showing utilisation of more than 100%. This can be true in
> some cases but for E1 interfaces I always thought that the router
> calculates the correct bw depending on the number of channels used. e.g
>
> router#sh run int s0/0:0
> Building configuration...
>
> Current configuration : 318 bytes
> !
> interface Serial0/0:0        <<no bandwidth configured>>
>  description ** To PE ***
>  no ip address
>  encapsulation frame-relay IETF
>  tx-ring-limit 2
>  tx-queue-limit 2
>  frame-relay lmi-type ansi
>  max-reserved-bandwidth 100
>  service-policy input IN-S0/0:0
>  service-policy output OUT-S0/0:0
> end
> !
> router#sh interface Serial0/0:0
> Serial0/0:0 is up, line protocol is up
>  Hardware is PowerQUICC Serial
>  description ** To PE ***
>  MTU 1500 bytes, BW 1984 Kbit, DLY 20000 usec,      <<bw 1984 kbps>>
>     reliability 255/255, txload 6/255, rxload 56/255
>  Encapsulation FRAME-RELAY IETF, loopback not set
> <<output omitted>>
>  Timeslot(s) Used:1-31, SCC: 0, Transmitter delay is 0 flags
> <<number of timeslots used>>
>
> But the bandwidth calculated for the sub-interface has a different
> value:
>
> rotuer#sh run int s0/0:0.101
> Building configuration...
>
> Current configuration : 175 bytes
> !
> interface Serial0/0:0.101 point-to-point      << also no bw statement>>
>  description Primary VPN WAN Link
>  ip unnumbered Loopback10
>  ip flow ingress
>  no cdp enable
>  frame-relay interface-dlci 101
>
> !
> rotuer#sh interface Serial0/0:0.101
> Serial0/0:0.101 is up, line protocol is up
>  Hardware is PowerQUICC Serial
>  Description: Primary VPN WAN Link
>  Interface is unnumbered. Using address of Loopback10
>  MTU 1500 bytes, BW 1024 Kbit, DLY 20000 usec,      <<bw 1024 kbps>>
>     reliability 255/255, txload 4/255, rxload 32/255
>  Encapsulation FRAME-RELAY IETF
>  Last clearing of "show interface" counters never
>
> Any ideas if this is a bug? Am I missing something here?
>
> Thanks in advance
>
>
> Nasir Shaikh

I would guess that your 31 channel E1 was upgraded sometime along the
way from a 16 channel service (16x64=1024).  The bandwidth of the
sub-interface was assigned when it was created and would not
dynamically adjust after more channels were assigned to the main
interface.

Andy
gawul00+cnsp at gmail.com