From: Ilya Mazhara (willy@xxxxxxxxxxxxxxxx)
Date: Thu Jan 18 2001 - 08:53:21 GMT-3
All RR in the same cluster share same cluster-id cause it works same
role as AS-tag in AS-path but
inside IBGP.
If u mean originator-id u're right, it should be unique at every RR.
Dave Humphrey wrote:
>
> Cisco recommend that each RR has a unique cluster-id. This info is contained
> in a presentation given at CCIE update session in Vienna.
>
> Dave Humphrey
>
> -----Original Message-----
> From: nobody@groupstudy.com [mailto:nobody@groupstudy.com]On Behalf Of
> tv
> Sent: 18 January 2001 04:45
> To: Daniel M. Dawson; Mark Stover
> Cc: ccielab@groupstudy.com
> Subject: Re: BGP Clusters with Multiple Route Reflectors
>
> EEK! I am sorry for that......I mis-read the post. I thought there were 2
> clusters. Yes, the cluster ids should be the same within a group of RRs
> that are providing redundancy to prevent routing loops. Going to bed now :)
>
> tv
> ----- Original Message -----
> From: "Mark Stover" <mstover@cisco.com>
> To: "'tv'" <tvarriale@telocity.com>; "'Daniel M. Dawson'"
> <dandawson@lucent.com>
> Cc: <ccielab@groupstudy.com>
> Sent: Wednesday, January 17, 2001 10:33 PM
> Subject: RE: BGP Clusters with Multiple Route Reflectors
>
> > No, All route-reflectors in a cluster (being used for redundancy) must
> have the same cluster-id, otherwise it is possible to get routing loops.
> >
> > > -----Original Message-----
> > > From: nobody@groupstudy.com [mailto:nobody@groupstudy.com]On Behalf Of
> > > tv
> > > Sent: Wednesday, January 17, 2001 11:21 PM
> > > To: Daniel M. Dawson
> > > Cc: ccielab@groupstudy.com
> > > Subject: Re: BGP Clusters with Multiple Route Reflectors
> > >
> > >
> > > The bgp cluster-ids MUST be different and only goes on the
> > > RRs..NOT the
> > > clients.
> > >
> > > tv
> > >
> > > ----- Original Message -----
> > > From: "Daniel M. Dawson" <dandawson@lucent.com>
> > > To: <ccielab@groupstudy.com>
> > > Sent: Wednesday, January 17, 2001 7:10 PM
> > > Subject: BGP Clusters with Multiple Route Reflectors
> > >
> > >
> > > > Hi all,
> > > >
> > > > A question about multiple route reflectors within a
> > > cluster. (I've set
> > > this
> > > > up based on Lab 5s layout). In this particular case I have
> > > R4 and R5 as
> > > the
> > > > route reflectors and R3 and R6 as the route reflector
> > > clients. These
> > > route
> > > > reflecotr clients are configured with neighbor statements
> > > to each route
> > > > reflector for redundancy. When I set up the
> > > route-reflectors only (no the
> > > > clients) with the cluster ID the clients get double listings for all
> > > > networks in their BGP table. The reflectors have only the
> > > single entry
> > > for
> > > > each entry. As shown below:
> > > >
> > > >
> > > > R5# "SHOW IP BGP" output (Route-reflector)
> > > > BGP table version is 5, local router ID is 172.168.200.1
> > > > Status codes: s suppressed, d damped, h history, * valid, >
> > > best, i -
> > > > internal
> > > > Origin codes: i - IGP, e - EGP, ? - incomplete
> > > >
> > > > Network Next Hop Metric LocPrf Weight Path
> > > > *>i20.0.0.0 172.168.100.3 0 100 0 ?
> > > > *> 71.0.0.0 137.20.20.1 0 2 3 i
> > > > *>i172.168.40.0/24 172.168.200.2 0 100 0 i
> > > > *>i172.168.60.0/24 172.168.100.6 0 100 0 i
> > > >
> > > > R6# "SHOW IP BGP" output (Route-reflector-client)
> > > > BGP table version is 17, local router ID is 172.168.100.6
> > > > Status codes: s suppressed, d damped, h history, * valid, >
> > > best, i -
> > > > internal
> > > > Origin codes: i - IGP, e - EGP, ? - incomplete
> > > >
> > > > Network Next Hop Metric LocPrf Weight Path
> > > > *>i20.0.0.0 172.168.100.3 0 100 0 ?
> > > > * i 172.168.100.3 0
> > > 100 0 ?
> > > > * i71.0.0.0 137.20.20.1 100
> > > 0 2 3 i
> > > > *>i 137.20.20.1
> > > 100 0 2 3 i
> > > > *>i172.168.40.0/24 172.168.200.2 0 100 0 i
> > > > * i 172.168.200.2 0
> > > 100 0 i
> > > > *> 172.168.60.0/24 0.0.0.0 0 32768 i
> > > >
> > > >
> > > > If instead I put the Cluster Id on the on the clients as
> > > well, the clients
> > > > have only the one entry for each network int the table.
> > > However a strange
> > > > thing occurs. Any routes coming from another client within
> > > the cluster
> > > does
> > > > not get put into the table of the other clients. The route
> > > reflectors
> > > > themselves do have the route in their BGP table and are
> > > advertizing it to
> > > > the clients. Show BGP table after putting clusterIDs on
> > > the clients shown
> > > > below:
> > > >
> > > >
> > > > R5# "SHOW IP BGP" output (Route-reflector)
> > > > BGP table version is 11, local router ID is 172.168.200.1
> > > > Status codes: s suppressed, d damped, h history, * valid, >
> > > best, i -
> > > > internal
> > > > Origin codes: i - IGP, e - EGP, ? - incomplete
> > > >
> > > > Network Next Hop Metric LocPrf Weight Path
> > > > *>i20.0.0.0 172.168.100.3 0 100 0 ?
> > > > *> 71.0.0.0 137.20.20.1 0 2 3 i
> > > > *>i172.168.40.0/24 172.168.200.2 0 100 0 i
> > > > *>i172.168.60.0/24 172.168.100.6 0 100 0 i
> > > >
> > > >
> > > >
> > > > R6# "SHOW IP BGP" output (Route-reflector-client)
> > > > BGP table version is 4, local router ID is 172.168.100.6
> > > > Status codes: s suppressed, d damped, h history, * valid, >
> > > best, i -
> > > > internal
> > > > Origin codes: i - IGP, e - EGP, ? - incomplete
> > > >
> > > > Network Next Hop Metric LocPrf Weight Path
> > > > *>i71.0.0.0 137.20.20.1 100 0 2 3 i
> > > > *>i172.168.40.0/24 172.168.200.2 0 100 0 i
> > > > *> 172.168.60.0/24 0.0.0.0 0 32768 i
> > > >
> > > >
> > > >
> > > > So the question is should I be using Cluster-ID on the
> > > > route-reflector-clients and if so, how do I get the client
> > > to see routes
> > > > originated on other clients. If I shouldn't be using
> > > cluster-ID on the
> > > > clients, are the double routes supposed to be there?
> > > Thanks for any help.
> > > >
> > > >
> > > > Configs:
> > > >
> > > > WITH CLUSTER ID ON THE ROUTE-REFLECTOR-CLIENT
> > > >
> > > > R4 (RR)
> > > > router bgp 1
> > > > no synchronization
> > > > bgp cluster-id 10
> > > > network 172.168.40.0 mask 255.255.255.0
> > > > neighbor 172.168.100.3 remote-as 1
> > > > neighbor 172.168.100.3 route-reflector-client
> > > > neighbor 172.168.100.5 remote-as 1
> > > > neighbor 172.168.100.6 remote-as 1
> > > > neighbor 172.168.100.6 route-reflector-client
> > > >
> > > > R3 (RRC)
> > > > router bgp 1
> > > > no synchronization
> > > > redistribute static
> > > > bgp cluster-id 10
> > > > neighbor 172.168.100.5 remote-as 1
> > > > neighbor 172.168.100.5 soft-reconfiguration inbound
> > > > neighbor 172.168.200.2 remote-as 1
> > > > neighbor 172.168.200.2 soft-reconfiguration inbound
> > > >
> > > >
> > > > R5 (RR)
> > > > router bgp 1
> > > > no synchronization
> > > > bgp cluster-id 10
> > > > neighbor 137.20.20.1 remote-as 2
> > > > neighbor 172.168.100.3 remote-as 1
> > > > neighbor 172.168.100.3 route-reflector-client
> > > > neighbor 172.168.100.6 remote-as 1
> > > > neighbor 172.168.100.6 route-reflector-client
> > > > neighbor 172.168.200.2 remote-as 1
> > > >
> > > >
> > > > R6 (RRC)
> > > > router bgp 1
> > > > no synchronization
> > > > bgp cluster-id 10
> > > > network 172.168.60.0 mask 255.255.255.0
> > > > neighbor 172.168.100.5 remote-as 1
> > > > neighbor 172.168.100.5 soft-reconfiguration inbound
> > > > neighbor 172.168.200.2 remote-as 1
> > > > neighbor 172.168.200.2 soft-reconfiguration inbound
> > > >
> > > >
> > > > WITHOUT CLUSTER ID ON THE ROUTE-REFLECTOR-CLIENT
> > > >
> > > > R4 (RR)
> > > > router bgp 1
> > > > no synchronization
> > > > bgp cluster-id 10
> > > > network 172.168.40.0 mask 255.255.255.0
> > > > neighbor 172.168.100.3 remote-as 1
> > > > neighbor 172.168.100.3 route-reflector-client
> > > > neighbor 172.168.100.5 remote-as 1
> > > > neighbor 172.168.100.6 remote-as 1
> > > > neighbor 172.168.100.6 route-reflector-client
> > > >
> > > > R3 (RRC)
> > > > router bgp 1
> > > > no synchronization
> > > > redistribute static
> > > > neighbor 172.168.100.5 remote-as 1
> > > > neighbor 172.168.100.5 soft-reconfiguration inbound
> > > > neighbor 172.168.200.2 remote-as 1
> > > > neighbor 172.168.200.2 soft-reconfiguration inbound
> > > >
> > > >
> > > > R5 (RR)
> > > > router bgp 1
> > > > no synchronization
> > > > bgp cluster-id 10
> > > > neighbor 137.20.20.1 remote-as 2
> > > > neighbor 172.168.100.3 remote-as 1
> > > > neighbor 172.168.100.3 route-reflector-client
> > > > neighbor 172.168.100.6 remote-as 1
> > > > neighbor 172.168.100.6 route-reflector-client
> > > > neighbor 172.168.200.2 remote-as 1
> > > >
> > > >
> > > > R6 (RRC)
> > > > router bgp 1
> > > > no synchronization
> > > > network 172.168.60.0 mask 255.255.255.0
> > > > neighbor 172.168.100.5 remote-as 1
> > > > neighbor 172.168.100.5 soft-reconfiguration inbound
> > > > neighbor 172.168.200.2 remote-as 1
> > > > neighbor 172.168.200.2 soft-reconfiguration inbound
> > > >
> > > >
> > > >
> > > >
> > > > Daniel M. Dawson
> > > > Network Systems Consultant
> > > > Network Care Professional Services (NPS)
> > > > Lucent Technologies
> > > > Mobile: (405) 850-1493
> > > > Pager: (800) 622-9202
> > > > E-mail: dandawson@lucent.com
> > > >
This archive was generated by hypermail 2.1.4 : Thu Jun 13 2002 - 10:27:34 GMT-3