From: Paul Borghese (pborghese@xxxxxxxxxxxxxx)
Date: Fri Jan 11 2002 - 20:34:54 GMT-3
When a route goes down, the router first checks for a feasable sucessor. If
there is no feasable sucessor, it then queries the directly connected
neighbors to see if they have a better route.
What is confusing is the term active. In most people vocabulary if you told
them that the route was active they would say "great". "Hey boss, I don't
know what they are complaining about the route is Active". In this case,
the router is actively looking for a new route :-)
Here is a paper by cisco that explains it far better then I.
http://www.cisco.com/warp/public/103/eigrp6.html
Paul
----- Original Message -----
From: "Huy Luu" <Huy_Luu%COMMONWEALTH@ccginc.com>
To: "Ellis, Neil" <Neil.Ellis@expanets.com>
Cc: <ccielab@groupstudy.com>
Sent: Friday, January 11, 2002 2:51 PM
Subject: RE: EIGRP routes
> Thankyou Neil
>
> I'm not sure if this info is correct, can you please verify
>
> For Eigrp, when a neighbor is unreachable after the hold time which is
> three times the hello, the route is actually mark active and remain so for
> a default of three minutes while DUAL finds an alternative route. So the
> process takes 3-4 minutes before the route is consider unreachable but
> still advertised as unreachable route to other live EIGRP neighbor.
>
> If you run OSPF and EIGRP and once the route is unavailable through EIGRP,
> how much time does it take for OSPF to take over and add an OSPF route to
> the routing table?
>
> Please confirm.
>
>
>
>
> "Ellis, Neil"
> <Neil.Ellis@exp To: 'Huy Luu'
<Huy_Luu%COMMONWEALTH@ccginc.com>,
> anets.com> ccielab@groupstudy.com
> cc: "Duffy, Bryan"
<Bryan.Duffy@expanets.com>
> 01/11/02 01:50 Subject: RE: EIGRP routes
> PM
>
>
>
>
>
>
>
> LUU:
>
>
>
> It depends on the routing protocol in use. OSPF uses a hello and dead
> interval. The hello depending upon the network type is default at 10
> seconds. The dead interval is always 4 times the hello interval. If a
> neighbor is not heard from by the time the dead interval is expired it is
> assumed down and taken out of the neighbor table. The hello interval can
be
> manually manipulate with the interface command ip ospf hello-interval.
>
http://www.cisco.com/univercd/cc/td/doc/product/software/ios121/121cgcr/ip_r
>
> /iprprt2/1rdospf.htm#xtocid796920
>
>
> At the time that the dead interval expires that router is no longer a
> neighbor and that whole routing table is taken out. The SPF tree also has
> to
> be recalculated. Good resources are Jeff Doyle TCP/IP Volume I and a new
> Cisco press book CCIE Practical Studies. Must reads for anyone attempting
> the CCIE Lab!
>
>
> With Eigrp neighbors are communicated with in a similar manner. Hellos are
> multicast out every 5 seconds or 60 seconds depending on the network type.
> It can be manipulated with the interface command ip hello-interval eigrp.
>
http://www.cisco.com/univercd/cc/td/doc/product/software/ios121/121cgcr/ip_r
>
> /iprprt2/1rdeigrp.htm#xtocid1895711
> The hello will include a hold time which is three times the hello timer.
> This determines how long to wait for another hello from the sending
> neighbor. Once that timer has expired the neighbor is considered gone. The
> default for hold time can be manipulated with the interface command ip
> hold-time eigrp.
>
http://www.cisco.com/univercd/cc/td/doc/product/software/ios121/121cgcr/ip_r
>
> /iprprt2/1rdeigrp.htm#xtocid1895712
> As with OSPF once the neighbor adjacency is lost all routes from that
> neighbor is cleared from the routing table.
>
>
> RIP has an expiration timer or timeout for routes in the routing table.
> Cisco may refer to it as an invalid timer. This is set at a default of 180
> seconds or 6 update periods. The protocol will change a route's hop count
> to
> 16 or unreachable if it has not been given an update for that particular
> route. Updates in the Cisco world can vary (according to Jeff Doyle's
> TCP/IP
> Volume I) between 25.5 to 30 seconds. Now there's another timer called the
> flush timer that is set to 240 seconds. The route is advertised as
> unreachable until that timer expires. At that time the route is finally
> removed from the router. The last timer is referred to as the holddown
> timer. Is is used when an update is received with a hop count greater than
> what exists in the current routing table. The protocol will place this
> route
> into holddown for 180 seconds.
>
>
> I hope all of this helps it is available in TCP/IP Volume I by Jeff Doyle.
> Good luck to all!
>
>
> Best Regards,
>
> Neil Ellis MCSE, CCNP,CCDP,A+,Net+,i-Net+,
> Network Engineer
> Converged Solutions
> * Neil.Ellis@expanets.com
> * 303-300-6320 x6224
> Cell 303-548-9453
>
>
>
> -----Original Message-----
> From: Huy Luu [mailto:Huy_Luu%COMMONWEALTH@ccginc.com]
> Sent: Friday, January 11, 2002 11:08 AM
> To: ccielab@groupstudy.com
> Subject: EIGRP routes
>
>
> When Eigrp, OSPF, RIP or any routing protocol looses a neighbor and does
> not have alternative route to networks advertised by the down neighbor,
> how long does it take for that neighbor's routes to dissapear from the
> routing table? Is there a default timeout value for routes in the
> routing table?Any link from CCO will be greatly appreciated. Thanks all
> for your help.
This archive was generated by hypermail 2.1.4 : Thu Jun 13 2002 - 10:56:25 GMT-3