From: William Dicks (wdicks@xxxxxxxxxxxxxxxxxx)
Date: Thu Sep 21 2000 - 15:05:06 GMT-3
Also remember that by default, the router will use the path that returns the
explorer packet first, whether the costs are different or not. (i.e. R1 peer
to R2 and R3. R2 has cost of 2 R3 has cost of 4. But if R3 returns the
explorer first, by default R1 will use R3). There is a global "dlsw timer
explorer-wait-time": (from doc cd) "Time to wait for all stations to respond
to explorers. The valid range is 1 to 86400 seconds. The default is 0."
(Since the default is 0, that is why the first response path is used. Just
make this 1 or 2 seconds in a test enviroment to see the explorers return
from all costed peers)
Bill Dicks
CCIE # 6081
-----Original Message-----
From: nobody@groupstudy.com [mailto:nobody@groupstudy.com]On Behalf Of
Mark Lewis
Sent: Thursday, September 21, 2000 12:05 PM
To: smaljure@cibernetworks.com; ccielab@groupstudy.com
Subject: Re: DLSw+ Backup peer v/s peers with different costs....
Nope. It's a little bit different.
With two peers with different costs, the 'local' router will of course use
(usually) the one with the lowest cost. However the other router will be in
a CONNECT state (capability exchange,etc. has taken place and keepalives are
being sent (assuming that keepalives are not set to 0)).
However, with a backup peer, the connection to the backup is not made until
there is a failure on the primary (so no traffic on backup path until then).
Hope this helps,
Mark
>From: smaljure@cibernetworks.com
>Reply-To: smaljure@cibernetworks.com
>CC: ccielab@groupstudy.com
>Subject: DLSw+ Backup peer v/s peers with different costs....
>Date: Thu, 21 Sep 2000 11:39:12 -0400
>
>
>Hi All,
>I was wondering if configuring a backup peer is the same as configuring two
>peers with different costs such that the lower cost peer is preferred and
>if
>the lower cost peer goes away, then the higher cost peer will be used...
>
>So, let us say RA has a DLSw peer relation with RB
>and RC is configured as a backup-peer on RA (to backup RB)
>
>v/s
>
>RA has two remote-peer statements for RB and RC and RB is configured with
>lower cost...
>and RC is configured with a higher cost
>
>Why would u prefer one method over the other?
>
>Thanks for your inputs
>Sanjay
>
This archive was generated by hypermail 2.1.4 : Thu Jun 13 2002 - 08:25:00 GMT-3