From: Jonathan V Hays (jhays@jtan.com)
Date: Fri Apr 11 2003 - 12:45:18 GMT-3
Page 673 of "CCIE Practical Studies, Volume I" by Karl Solie:
"The delay is the summation of the delays on outbound interfaces only."
Page 675:
Be sure to include a day at each side of the interface if you want
symmetrical routing - that is, packets will take the same route back to
the source."
In my own experience, there are unfortunately a number of IOS changes
that will that take place in their own sweet time (slowly) so you need
to "clear ip eigrp neighbors" or even reboot the router.
> -----Original Message-----
> From: nobody@groupstudy.com [mailto:nobody@groupstudy.com] On
> Behalf Of Ram Shummoogum
> Sent: Friday, April 11, 2003 7:50 AM
> To: cash2001@swbell.net
> Cc: ccielab@groupstudy.com
> Subject: Re: EiGRP metric - delay on serial
>
>
> Jason:
>
> The feasible distance is equal to the advertised distance
> (or reported distance , depending which book you read) by R6
> + the metric it takes to reach R6 .
>
> In other words, changing the delay on the serial of r6 will
> have no impact on the metric. What you have to do is change
> in on the serial of r2.
>
> But obviously in your case the load balancing will still
> occur even if you do the change on the serial of r2 as it is
> a multipoint link to r5 and r6.
>
> In order to make r2 prefer r5 in your case would be to change
> the delay on R6 ethernet interface( example on the 150.50.7.0 subnet)
>
>
>
> HTH
>
> RAM
>
>
>
>
>
> "Jason Cash" <cash2001@swbell.net>@groupstudy.com on
> 04/11/2003 12:29:12 AM
>
> Please respond to "Jason Cash" <cash2001@swbell.net>
>
> Sent by: nobody@groupstudy.com
>
>
> To: <ccielab@groupstudy.com>
> cc:
> Subject: EiGRP metric - delay on serial
>
>
> I am trying to alter the metrics of a route in the IPE Lab 7.
> R2 is rcv'ing routes from R5 and R6 via serial links. I
> change the delay on the serial interfaces of R6 to have R2
> prefer the route to R5, but it has no effect:
>
> r2#si
>
> 150.50.0.0/16 is variably subnetted, 7 subnets, 4 masks
> D 150.50.4.0/24 [90/2733056] via 150.50.100.5, 00:00:07,
> Serial0.256 R5
> [90/2733056] via 150.50.100.6, 00:00:08,
> Serial0.256 R6
> D 150.50.7.0/25 [90/2195456] via 150.50.100.5, 00:00:18,
> Serial0.256
> [90/2195456] via 150.50.100.6,
> 00:00:18, Serial0.256
> D 150.50.5.64/27 [90/2707456] via 150.50.100.5, 00:00:08,
> Serial0.256
> [90/2707456] via 150.50.100.6,
> 00:00:08, Serial0.256
>
> Then change the delay on the serial link on R6:
>
> r6#conf t
> Enter configuration commands, one per line. End with CNTL/Z.
> r6(config)#int s1/0 r6(config-if)#delay 100000
>
> r2#clear ip ro *
> r2#si
>
> 150.50.0.0/16 is variably subnetted, 7 subnets, 4 masks
> D 150.50.4.0/24 [90/2733056] via 150.50.100.6, 00:00:01,
> Serial0.256
> [90/2733056] via 150.50.100.5,
> 00:00:02, Serial0.256
> D 150.50.7.0/25 [90/2195456] via 150.50.100.6, 00:00:02,
> Serial0.256
> [90/2195456] via 150.50.100.5,
> 00:00:02, Serial0.256
> D 150.50.5.64/27 [90/2707456] via 150.50.100.6, 00:00:02,
> Serial0.256
> [90/2707456] via 150.50.100.5,
> 00:00:02, Serial0.256
>
> Changing the serial interface delay on S1/0 of R6 to had no
> affect! IS something wrong here?
>
> Serial1/0 is up, line protocol is up
> Hardware is cxBus Serial
> Internet address is 150.50.100.6/27
> MTU 1500 bytes, BW 1544 Kbit, DLY 1000000 usec,
>
>
> Now when I change the delay on the Ethernet interface (ingress path of
> route) the metric is changed, but I thought all paths were
> considered when tallying delay.
This archive was generated by hypermail 2.1.4 : Thu May 01 2003 - 13:35:51 GMT-3