From: kevin gannon (kevin@gannons.net)
Date: Tue Oct 18 2005 - 04:20:03 GMT-3
Paresh,
I missed the difference in metric size between inter and intra area.
I retested and it worked as expected thanks for clearing that up
for me.
Regards
Kevin
On 10/18/05, Andrew Lissitz (alissitz) <alissitz@cisco.com> wrote:
> Understood and thanks for clarifying about the "only path to
> destination"... this actually makes sense... ;-)
>
> I went back and reread your earlier response and understand better of
> what you meant. I had originally thought that you meant that 16 bits
> are not enough to influence OSPF path selection ... Good stuff Paresh,
>
> Andrew
>
>
>
> -----Original Message-----
> From: nobody@groupstudy.com [mailto:nobody@groupstudy.com] On Behalf Of
> Paresh Khatri
> Sent: Monday, October 17, 2005 7:01 PM
> To: Andrew Lissitz (alissitz); kevin gannon; Ccie Lab (E-mail)
> Cc: C&S GroupStudy
> Subject: RE: ospf max-mteric router
>
> Hi Andrew,
>
> What max-metric will achieve is to ensure that this router will not be
> used for *transit* traffic. However, all stub routes injected by this
> router will still be reachable from other routers. According to
> RFC2328, a metric of 65535 for a link is *not* considered unreachable.
> Therefore, if this router is the only way to reach a particular
> destination, it will still be used without any problems.
>
> Paresh.
>
> -----Original Message-----
> From: Andrew Lissitz (alissitz) [mailto:alissitz@cisco.com]
> Sent: Tuesday, 18 October 2005 08:55 AM
> To: Paresh Khatri; kevin gannon; Ccie Lab (E-mail)
> Cc: C&S GroupStudy
> Subject: RE: ospf max-mteric router
>
>
> I think I may be missing something here but ...
>
> If an OSPF router advertises it's own reachability (LSA-1) as maxed ...
> max-metric ... then what other OSPF router will prefer this router?
>
> As each router constructs a SPF tree, each router sees the network from
> their own perspective (they are the base of the tree) and this router
> will not look appealing, especially the closer you get to it.
>
> When this router's networks are advertised across the OSPF network then
> just about any other path will be good. The closer the routers are to
> this router, the worse it looks, again this is because SPF is built on
> local perspective.
>
> Now if this is the only way to reach a destination ... then maxing a
> metric will not help.
>
> Please correct me if I am not in tune with what you are trying to do /
> achieve. Kindest regards
>
>
>
> -----Original Message-----
> From: nobody@groupstudy.com [mailto:nobody@groupstudy.com] On Behalf Of
> Paresh Khatri
> Sent: Monday, October 17, 2005 6:39 PM
> To: kevin gannon; Ccie Lab (E-mail)
> Subject: RE: ospf max-mteric router
>
> Hi Kevin,
>
> According to RFC2328, the metric value in type-1 (router-LSAs) LSAs is
> 16-bits, while the metric value in type-3, type-4 and type-5 LSAs is
> 24bits. Therefore, the maximum metric value for type-1 LSAs is 65535
> while that for the other LSA types is 16777215.
>
> The metric that you specify with the 'ip ospf cost' command is that for
> a router link, which appears in the router-LSAs and therefore it is a
> maximum of 65535. The metric indicated in your output is for an
> inter-area route, which can be up to 24 bits.
>
> There is a good reason why the metric for the inter-area routes is 24
> bits - when you add up 16-bit metric values, it is quite possible that
> you will get a value greater than 16 bits. Since the metric for
> inter-area routes is a summation of the metrics of individual links that
> make up the path, we need a metric greater than 16-bits. The OSPF
> developers chose to use 24-bits.
>
> Hope that helps,
> Paresh Khatri
>
> -----Original Message-----
> From: nobody@groupstudy.com [mailto:nobody@groupstudy.com]On Behalf Of
> kevin gannon
> Sent: Tuesday, 18 October 2005 03:16 AM
> To: Ccie Lab (E-mail)
> Subject: ospf max-mteric router
>
>
> Just looking at this command again and you would think that it sets the
> maximum possible metric for the routes. It states it sets them to 65535.
> This is not the max metric as you can see from below:
>
> SW2(config-if)#do sh ip rou 2.2.2.2
> Routing entry for 2.2.2.2/32
> Known via "ospf 1", distance 110, metric 131071, type inter area
> Last update from 14.14.14.1 on FastEthernet0/14, 00:00:02 ago
> Routing Descriptor Blocks:
> * 14.14.14.1, from 123.123.123.2, 00:00:02 ago, via FastEthernet0/14
> Route metric is 131071, traffic share count is 1
>
> SW2(config-if)#
>
> The metric is 131071. Any ideas why this might be the case, I know its
> highly unlikey to get a cost higher that
> 65535 in real life. Its just someone in Cisco thought up this value
> maybe its a safety mechanism just in case the router might be able to
> route packets ?
>
> Regards
> Kevin
>
> _______________________________________________________________________
> Subscription information may be found at:
> http://www.groupstudy.com/list/CCIELab.html
>
>
> This communication, including any attachments, is confidential. If you
> are not the intended recipient, you should not read it - please contact
> me immediately, destroy it, and do not copy or use any part of this
> communication or disclose anything about it.
>
> _______________________________________________________________________
> Subscription information may be found at:
> http://www.groupstudy.com/list/CCIELab.html
>
>
> This communication, including any attachments, is confidential. If you
> are not the intended recipient, you should not read it - please contact
> me immediately, destroy it, and do not copy or use any part of this
> communication or disclose anything about it.
>
> _______________________________________________________________________
> Subscription information may be found at:
> http://www.groupstudy.com/list/CCIELab.html
This archive was generated by hypermail 2.1.4 : Sun Nov 06 2005 - 22:00:51 GMT-3