From: Ramiro Garza (ccie15xxx@gmail.com)
Date: Mon Dec 12 2005 - 00:43:39 GMT-3
IMHO Godswill is absolutely right, getting back to the original question,
you can do it whichever way you want unless told otherwise, example follows.
let's say you have a router with only 3 interfaces and with the following ip
addressess and you want to run ospf on all of them and on the same area.
int eth 1
ip add 10.1.1.1 255.255.255.0
int eth 2
ip add 10.1.2.1 255.255.255.0
int eth 3
ip add 10.1.3.1 255.255.255.0
what would be easier and faster to configure?
router ospf 1
10.1.1.1 0.0.0.0 area x
10.1.2.1 0.0.0.0 area x
10.1.3.1 0.0.0.0 area x
or
router ospf 1
10.1.0.0 0.0.255.255 area x
Hope this clears up the first question.
For the second question, ALWAYS configure the router-id under the OSPF
process (unless told otherwise)
PS For labs I always use 0.0.0.0
Regards.
On 12/9/05, Godswill Oletu <oletu@inbox.lv> wrote:
>
> James,
>
> You are mistaking the native behaviour of the various ospf network types
> with the function of the network command. Lab up your solution and see
> what
> you get.
>
> Natively, point-to-multipoint & point-to-multipoint non-broadcast ospf
> network types will advertise their links as a host routes (/32) to their
> neighbor and this is inrespective of the mask you used to advertise those
> interfaces.
>
> Broadcast, non-broadcast & point-to-point ospf network types will
> advertise
> their links as a prefix, this again works independent of the mask value
> used
> to advertise them.
>
> The mask value in the ospf network statement, just scan through your
> router
> and add all matching interfaces into the specified ospf area, after that,
> the router will advertise the prefix/ip addresses configured on those
> interfaces. How these prefix/ip addresses are advertised, is greatly
> influenced by the ospf network type configured for that interface.
>
> If one's scanario prevent him/her from using the ospf network type
> statement
> to change the mask of the advertised network, there are tons of other ways
> depending on other restrictions in place, to do that.
>
> HTH
> Godswill
>
>
> ----- Original Message -----
> From: "James Matrisciano" <jmatrisciano@kenttech.com>
> To: "Schulz, Dave" <DSchulz@dpsciences.com>; "Melwani, Manoj J"
> <melwanim@citigroup.com>; "Kim Judy" <kim23uk@gmail.com>;
> <ccielab@groupstudy.com>
> Sent: Friday, December 09, 2005 5:48 PM
> Subject: RE: Wildcard issue
>
>
> > With the following
> >
> > Ospf router 1
> > Net 1.1.1.1 0.0.0.0 area 0
> > Net 3.3.3.1 0.0.0.0 area 0
> >
> > Int s0/0
> > Ip add 1.1.1.1 255.255.255.0
> > Ip ospf net broadcast
> >
> > Int s0/0.3 m
> > Ip add 3.3.3.1 255.255.255.0
> > Ip ospf net point-to-multipoint
> >
> > You area connected to router 2 and router 3 via Frame all DLCI's are
> > mapped
> >
> > Once your network has converged, you will get statements in your routing
> > table with a /32 of the interface that router is connected to. This
> > will break the restriction of advertise the interface with its configure
> > mask.
> >
> > With this set up (and forgive me, my lab is not on line, so I can not
> > copy paste my sh ip route statements)
> >
> > In router 3 that is connected to router 1 with
> > S2/0
> > Ip add 3.3.3.3 255.255.255.0
> > ip ospf net point-to-multipoint
> >
> > Router ospf 1
> > Net 3.3.3.3 0.0.0.0 area 0
> >
> > Sh ip route
> >
> > 1.0.0.0
> > O 1.1.1.0 (these are the routes to get to R2
> > 3.0.0.0/8
> > O 3.3.3.1/32
> > C 3.3.0/24
> >
> > What you want to look for is that if you have a restriction saying
> > "advertise as configured" the 3.3.3.1 interface is configured as a /24,
> > not a /32
> >
> > As well, in router 2 that is connected the same way except with the ip
> > ospf network broadcast command on the interface, the routing table
> > shows:
> >
> > 3.0.0.0/32
> >
> > In its routing table, this would be a break in the restriction.
> > Using ip ospf network point-to-point fixes this problem, but only if you
> > are allowed to have point-to-point neighbor relationships
> >
> >
> > -----Original Message-----
> > From: Schulz, Dave [mailto:DSchulz@dpsciences.com]
> > Sent: Friday, December 09, 2005 5:33 PM
> > To: James Matrisciano; Melwani, Manoj J; Kim Judy;
> > ccielab@groupstudy.com
> > Subject: RE: Wildcard issue
> >
> > James -
> >
> > Can you please expound on the advertisement issue that you state. I
> > didn't think that the network command under ospf had anything to do with
> > how the interface was advertised. I thought that this only determined
> > which interfaces were included in the ospf process. I do agree that the
> > ospf network command does affect the advertisement.
> >
> >
> > Dave Schulz
> > Email: dschulz@dpsciences.com
> >
> >
> >
> > -----Original Message-----
> > From: nobody@groupstudy.com [mailto:nobody@groupstudy.com] On Behalf Of
> > James Matrisciano
> > Sent: Friday, December 09, 2005 5:13 PM
> > To: Melwani, Manoj J; Kim Judy; ccielab@groupstudy.com
> > Subject: RE: Wildcard issue
> >
> > You need to watch out for your restrictions on the 0.0.0.0 mask. If you
> > have a restriction that all interfaces must be advertised with their
> > original mask, make sure you do not use the 0.0.0.0 unless you are
> > allowed to use the ip ospf network point-to-point command under the
> > interface you are conneting with. Again, this will fall under the
> > restrictions of how you can have your ospf neighbors talk to each other.
> >
> >
> > -----Original Message-----
> > From: nobody@groupstudy.com [mailto:nobody@groupstudy.com] On Behalf Of
> > Melwani, Manoj J
> > Sent: Friday, December 09, 2005 4:45 PM
> > To: Kim Judy; ccielab@groupstudy.com
> > Subject: RE: Wildcard issue
> >
> > The best practice is to use the second method. By using 0.0.0.0 wildcard
> > you are saying that I want to run OSPF routing process just on that
> > particular interface unless if you have a whole range of interface that
> > you want to use on the same line then you might want to use the first
> > method.
> >
> > Yes Its a good idea to use router-id under the routing process. Use
> > Loopback 0 as your router id's.
> >
> >
> > Thanks,
> > Manoj.
> >
> > -----Original Message-----
> > From: nobody@groupstudy.com [mailto:nobody@groupstudy.com]On Behalf Of
> > Kim Judy
> > Sent: Friday, December 09, 2005 4:15 PM
> > To: ccielab@groupstudy.com
> > Subject: Wildcard issue
> >
> >
> > hi group,
> > I have a query that if a interface has an Ip address 172.16.12.1/25 What
> > wild card bits should I use under the IGP process ?
> >
> > router ospf 1
> > net 172.16.12.1 0.0.0.127 or
> > net 172.16.12.1 0.0.0.0
> >
> > which is the best way ? What does Cisco expect from us ?
> >
> > If not specified should I use a "router-id" under each IGP running on a
> > router ?
> >
> > _______________________________________________________________________
> > Subscription information may be found at:
> > http://www.groupstudy.com/list/CCIELab.html
> >
> > _______________________________________________________________________
> > Subscription information may be found at:
> > http://www.groupstudy.com/list/CCIELab.html
> >
> > _______________________________________________________________________
> > Subscription information may be found at:
> > http://www.groupstudy.com/list/CCIELab.html
> >
> > _______________________________________________________________________
> > Subscription information may be found at:
> > http://www.groupstudy.com/list/CCIELab.html
>
> _______________________________________________________________________
> Subscription information may be found at:
> http://www.groupstudy.com/list/CCIELab.html
This archive was generated by hypermail 2.1.4 : Mon Jan 09 2006 - 07:07:50 GMT-3