From: George Zhang (gyzhang@xxxxxxxxxxx)
Date: Sun Dec 03 2000 - 02:02:01 GMT-3
Chuck,
You are right about the fact that after the local router RID is changed but the
old RID is still in neighbor router's database. I guess that it is due to agin
g
process of entries in the OSPF database. Change of local RID is not reflected
in
the neighbor immediately. Rebooting the router that has its RID changed may no
t
refresh the neighbor router's database either. When I was playing in the lab a
while ago, I ran into a situation that has three router (A ,B, and C) connected
in full mesh configured with OSPF. I changed the RID of Router A. Then, I
rebooted router A. After that, the old RID of Router A is still in a router B
database. I rebooted Router B, but I still saw old RID of Router A in it's
database after it back up. So I thought that old RID of router A in router B's
database after router B was reboot was heard from router C. So I rebooted rout
er
C. However, that did not fixed it either. My last resort was to reboot all
three router TOGETHER. That finally fixed the problem. I wonder any one else
had similar experience.
So I think that a sure way to change RID of a router and have that change
reflected immediately in neighbor router's database is to reboot that router an
d
all it's directly connected neighbors. Reconfiguring OSPF might fix the proble
m
too.
I hope it helps.
George Zhang
Chuck Larrieu wrote:
> It may be true that you have successfully changed one router i.d. by this
> process. However, my experiments continue to indicate that old or bad RIDs
> continue to exist on other routers.
>
> Router1 - adjacent to router with RID 10.5.0.2
> Router2 - adjacent to router with RID 10.5.0.1
>
> I have made several changes to router 2 IP addresses on various interfaces.
> I have done shutdowns and no shuts on interfaces. I have blown away
> loopbacks and restored them with new IP addresses.
>
> Now router1 sh ip protocol indicates it is receiving routes from routers
> with RID's 10.5.0.2, and 10.7.0.2 ( changes I have made at one time or
> another during this discussion ) although Router1 show IP ospf neighbor
> indicates it has an adjacency with Router2 with a RID of 10.2.0.2 ( the
> highest IP after the last of my changes ), the show ip protocol does not
> show as a source 10.2.0.2, it still shows 10.7.0.2 and 10.5.0.2, neither of
> which addresses exist. Clear ip ospf process does not change this. Only a
> reload seems to do this.
>
> Someone in another thread was asking about pitfalls. I believe this
> discussion has revealed a pretty big pitfall in the OSPF process.
>
> Chuck
>
> -----Original Message-----
> From: nobody@groupstudy.com [mailto:nobody@groupstudy.com] On Behalf Of Wu,
> Jiang
> Sent: Saturday, December 02, 2000 2:48 AM
> To: erickbe@yahoo.com
> Cc: ccielab@groupstudy.com
> Subject: Re: OSPF Lab - RID behavior
>
> It works on 11.3(11a)T1 but not on 12.0(13). Seems to be a dated function?
>
> Wu
>
> ----- Original Message -----
> From: Erick B. <erickbe@yahoo.com>
> To: George Zhang <gyzhang@bigfoot.com>; Chuck Larrieu <chuck@cl.cncdsl.com>
> Cc: Wu, Jiang <wujiang@bj163.com>; <ccielab@groupstudy.com>
> Sent: Saturday, December 02, 2000 6:13 PM
> Subject: Re: OSPF Lab - RID behavior
>
> > This works as well. Just tried it on 12.0(5) Mainline
> >
> > All you need to do is shutdown the interface. OSPF
> > will start using the highest active IP address
> > automatically or if you shutdown all your IP
> > interfaces, it spits out this error over and over
> > again which is expected.
> >
> > 2w2d: %OSPF-4-NORTRID: Could not allocate router ID
> >
> > - Erick
> >
> > 24d: %VOTE-2000: Counting error in FL, USA
> >
> > --- George Zhang <gyzhang@bigfoot.com> wrote:
> > > If you configured OSPF before you configured an loop
> > > back interface, the RID will be
> > > the ip address of a physical interface. Then, after
> > > you configuring an loopback
> > > interface, you can do shut/no shut on the physical
> > > interface that has the RID as its
> > > ip address. By doing so, the RID will be changed to
> > > the ip address of the loopback
> > > interface. At least, it worked for me the other day
> > > (with IOS 11.3). Correct me if I
> > > am wrong.
> > >
> > > George Zhang
> > >
> > > Chuck Larrieu wrote:
> > >
> > > > Jiang, I'm running IOS 12.1 in my lab. Clear IP
> > > OSPF process does not clear
> > > > out bad or old RID information
> > > >
> > > > Various experiments over the last couple of days,
> > > both as a result of this
> > > > thread and private conversations, led to my
> > > discovery that blowing out the
> > > > OSPF configuration completely, then rebuilding it,
> > > will remove the bad
> > > > information ( such as a bad neighboring RID ). And
> > > so will reloading the
> > > > router. But once RID's are learned, at least in
> > > what I have seen so far, it
> > > > is hell getting "rid" of them.
> > > >
> > > > Chuck
> > > >
> > > > -----Original Message-----
> > > > From: nobody@groupstudy.com
> > > [mailto:nobody@groupstudy.com] On Behalf Of Wu,
> > > > Jiang
> > > > Sent: Friday, December 01, 2000 9:00 PM
> > > > To: ccielab@groupstudy.com
> > > > Subject: Re: OSPF Lab - DR behaviour with
> > > loopbacks WAS: RE: question about
> > > > loopback interfaces
> > > >
> > > > In some ios versions (maybe 12.0 GD), you can use
> > > "clear ip ospf process"
> > > > command to restart OSPF.
> > > >
> > > > Wu
> > > >
> > > > ----- Original Message -----
> > > > From: Tony Olzak <aolzak@buckeye-express.com>
> > > > To: <erickbe@yahoo.com>; Chuck Larrieu
> > > <chuck@cl.cncdsl.com>; Louie Belt
> > > > <louieb@netmatter.com>; 'CCIE_Lab Groupstudy List'
> > > <ccielab@groupstudy.com>;
> > > > <cisco@groupstudy.com>
> > > > Sent: Saturday, December 02, 2000 5:16 AM
> > > > Subject: Re: OSPF Lab - DR behaviour with
> > > loopbacks WAS: RE: question about
> > > > loopback interfaces
> > > >
> > > > > I usually just reboot routers on the fly and
> > > work on something else while
> > > > > that router is rebooting.
> > > > >
> > > > >
> > > > > Tony
> > > > >
> > > > > ----- Original Message -----
> > > > > From: "Erick B." <erickbe@yahoo.com>
> > > > > To: "Tony Olzak" <aolzak@buckeye-express.com>;
> > > "Chuck Larrieu"
> > > > > <chuck@cl.cncdsl.com>; "Louie Belt"
> > > <louieb@netmatter.com>; "'CCIE_Lab
> > > > > Groupstudy List'" <ccielab@groupstudy.com>;
> > > <cisco@groupstudy.com>
> > > > > Sent: Friday, December 01, 2000 3:35 PM
> > > > > Subject: Re: OSPF Lab - DR behaviour with
> > > loopbacks WAS: RE: question
> > > > about
> > > > > loopback interfaces
> > > > >
> > > > >
> > > > > > If you remove the router ospf configuration
> > > and paste
> > > > > > it back, OSPF will restart with a new router
> > > ID if you
> > > > > > have a new high IP address. You can only do
> > > this in a
> > > > > > test/non-production network environment
> > > though. I've
> > > > > > done this before in my labs because it is
> > > faster then
> > > > > > waiting for the router to reboot.
> > > > > >
> > > > > > > And you are right, the RID doesn't change at
> > > all
> > > > > > > without rebooting the
> > > > > > > router. But, what do most techs do when a
> > > link is
> > > > > > > having problems? Reboot
> > > > > > > the routers. Now your RID will change.
> > > > > > >
> > > > > > > Tony
> >
> >
> >
This archive was generated by hypermail 2.1.4 : Thu Jun 13 2002 - 08:25:57 GMT-3