Mailing List Archive: Ibgp Routes.

Ibgp Routes.

Mar 6, 2006, 2:18 PM

Post #1 of 4 (1118 views)

Hi All,

Does anyone has an explination for this?
- Two routers in VRRP-E.
- IBGP session between them.

Why don't the IBGP peers sent eachother their entire received full-transtit
routes?
Does not make sense to me :-(

Thanks in advance.

Regards,
Armand

Summary:
+++++++++++++++++++++++++++++++++++++++++
RT1:
SSH at ams-gs-rt1>show ip bgp summary
BGP4 Summary
Router ID: 195.144.2.2 Local AS Number : 35829
Confederation Identifier : not configured
Confederation Peers:
Maximum Number of Paths Supported for Load Sharing : 1
Number of Neighbors Configured : 2, UP: 2
Number of Routes Installed : 325556
Number of Routes Advertising to All Neighbors : 118348
Number of Attribute Entries Installed : 61441
Neighbor Address AS# State Time Rt:Accepted Filtered Sent
ToSend
195.144.2.3 35829 ESTAB 3d 9h59m 143228 0 118347 0
217.170.9.129 24785 ESTAB 4d12h11m 182327 0 1 0

RT2:
SSH at ams-gs-rt2>show ip bgp summary
BGP4 Summary
Router ID: 84.233.188.98 Local AS Number : 35829
Confederation Identifier : not configured
Confederation Peers:
Maximum Number of Paths Supported for Load Sharing : 1
Number of Neighbors Configured : 2, UP: 2
Number of Routes Installed : 297731
Number of Routes Advertising to All Neighbors : 143226
Number of Attribute Entries Installed : 52427
Neighbor Address AS# State Time Rt:Accepted Filtered Sent
ToSend
84.233.188.97 8928 ESTAB 7d 2h57m 179398 0 1 0
195.144.2.2 35829 ESTAB 3d 9h58m 118332 0 143225 0

Ibgp Routes. [ In reply to ]

steve at telecomplete

Mar 6, 2006, 2:31 PM

Post #2 of 4 (1096 views)

Permalink

thats normal bgp behaviour

ibgp does not send full routes unless you configure them to each be route
reflector clients.. they will only send the best external routes

Steve

On Mon, 6 Mar 2006, A. Krijgsman - QCC wrote:

> Hi All,
>
> Does anyone has an explination for this?
> - Two routers in VRRP-E.
> - IBGP session between them.
>
> Why don't the IBGP peers sent eachother their entire received full-transtit
> routes?
> Does not make sense to me :-(
>
> Thanks in advance.
>
> Regards,
> Armand
>
>
> Summary:
> +++++++++++++++++++++++++++++++++++++++++
> RT1:
> SSH at ams-gs-rt1>show ip bgp summary
> BGP4 Summary
> Router ID: 195.144.2.2 Local AS Number : 35829
> Confederation Identifier : not configured
> Confederation Peers:
> Maximum Number of Paths Supported for Load Sharing : 1
> Number of Neighbors Configured : 2, UP: 2
> Number of Routes Installed : 325556
> Number of Routes Advertising to All Neighbors : 118348
> Number of Attribute Entries Installed : 61441
> Neighbor Address AS# State Time Rt:Accepted Filtered Sent
> ToSend
> 195.144.2.3 35829 ESTAB 3d 9h59m 143228 0 118347 0
> 217.170.9.129 24785 ESTAB 4d12h11m 182327 0 1 0
>
>
> RT2:
> SSH at ams-gs-rt2>show ip bgp summary
> BGP4 Summary
> Router ID: 84.233.188.98 Local AS Number : 35829
> Confederation Identifier : not configured
> Confederation Peers:
> Maximum Number of Paths Supported for Load Sharing : 1
> Number of Neighbors Configured : 2, UP: 2
> Number of Routes Installed : 297731
> Number of Routes Advertising to All Neighbors : 143226
> Number of Attribute Entries Installed : 52427
> Neighbor Address AS# State Time Rt:Accepted Filtered Sent
> ToSend
> 84.233.188.97 8928 ESTAB 7d 2h57m 179398 0 1 0
> 195.144.2.2 35829 ESTAB 3d 9h58m 118332 0 143225 0
>
>
>
>
> _______________________________________________
> foundry-nsp mailing list
> foundry-nsp at puck.nether.net
> http://puck.nether.net/mailman/listinfo/foundry-nsp
>

Ibgp Routes. [ In reply to ]

a.krijgsman at quality-control-centre

Mar 7, 2006, 1:03 AM

Post #3 of 4 (1105 views)

Permalink

Dear Steven,

As you can see in the Summary, both routes have only one transit uplink.
Each router only has one bgp session for external routes.
Wouldn't that mean that they need to sent all eBGP routes to the iBGP
neighbor?
Or does it match the routes from both routers?

I wonder what happens if one of our uplinks fails. Will I be able to sent
full transit over the iBGP session?

Regards,
Armand

----- Original Message -----
From: "Stephen J. Wilcox" <steve@telecomplete.co.uk>
To: "A. Krijgsman - QCC" <a.krijgsman at quality-control-centre.net>
Cc: <foundry-nsp at puck.nether.net>
Sent: Monday, March 06, 2006 11:31 PM
Subject: Re: [f-nsp] Ibgp Routes.

>
> thats normal bgp behaviour
>
> ibgp does not send full routes unless you configure them to each be route
> reflector clients.. they will only send the best external routes
>
> Steve
>
> On Mon, 6 Mar 2006, A. Krijgsman - QCC wrote:
>
>> Hi All,
>>
>> Does anyone has an explination for this?
>> - Two routers in VRRP-E.
>> - IBGP session between them.
>>
>> Why don't the IBGP peers sent eachother their entire received
>> full-transtit
>> routes?
>> Does not make sense to me :-(
>>
>> Thanks in advance.
>>
>> Regards,
>> Armand
>>
>>
>> Summary:
>> +++++++++++++++++++++++++++++++++++++++++
>> RT1:
>> SSH at ams-gs-rt1>show ip bgp summary
>> BGP4 Summary
>> Router ID: 195.144.2.2 Local AS Number : 35829
>> Confederation Identifier : not configured
>> Confederation Peers:
>> Maximum Number of Paths Supported for Load Sharing : 1
>> Number of Neighbors Configured : 2, UP: 2
>> Number of Routes Installed : 325556
>> Number of Routes Advertising to All Neighbors : 118348
>> Number of Attribute Entries Installed : 61441
>> Neighbor Address AS# State Time Rt:Accepted Filtered Sent
>> ToSend
>> 195.144.2.3 35829 ESTAB 3d 9h59m 143228 0 118347 0
>> 217.170.9.129 24785 ESTAB 4d12h11m 182327 0 1 0
>>
>>
>> RT2:
>> SSH at ams-gs-rt2>show ip bgp summary
>> BGP4 Summary
>> Router ID: 84.233.188.98 Local AS Number : 35829
>> Confederation Identifier : not configured
>> Confederation Peers:
>> Maximum Number of Paths Supported for Load Sharing : 1
>> Number of Neighbors Configured : 2, UP: 2
>> Number of Routes Installed : 297731
>> Number of Routes Advertising to All Neighbors : 143226
>> Number of Attribute Entries Installed : 52427
>> Neighbor Address AS# State Time Rt:Accepted Filtered Sent
>> ToSend
>> 84.233.188.97 8928 ESTAB 7d 2h57m 179398 0 1 0
>> 195.144.2.2 35829 ESTAB 3d 9h58m 118332 0 143225 0
>>
>>
>>
>>
>> _______________________________________________
>> foundry-nsp mailing list
>> foundry-nsp at puck.nether.net
>> http://puck.nether.net/mailman/listinfo/foundry-nsp
>>
>
>

Ibgp Routes. [ In reply to ]

steve at telecomplete

Mar 7, 2006, 2:48 AM

Post #4 of 4 (1119 views)

Permalink

Hi Armand,
no, they just send the best ebgp routes to each other - therefore each router
has a copy of the full routing table based on all the routes from its ebgp
neighbors and the (best) routes from the ibgp neighbors.

If one of the transits fails then the other router will have all the best routes
and will advertise them all to its ibgp neighbor. The convergence will take a
couple of minutes if we assume thats going to be about 90000 routes.

Steve

On Tue, 7 Mar 2006, A.P. Krijgsman wrote:

> Dear Steven,
>
> As you can see in the Summary, both routes have only one transit uplink.
> Each router only has one bgp session for external routes.
> Wouldn't that mean that they need to sent all eBGP routes to the iBGP
> neighbor?
> Or does it match the routes from both routers?
>
> I wonder what happens if one of our uplinks fails. Will I be able to sent
> full transit over the iBGP session?
>
> Regards,
> Armand
>
>
> ----- Original Message -----
> From: "Stephen J. Wilcox" <steve at telecomplete.co.uk>
> To: "A. Krijgsman - QCC" <a.krijgsman at quality-control-centre.net>
> Cc: <foundry-nsp at puck.nether.net>
> Sent: Monday, March 06, 2006 11:31 PM
> Subject: Re: [f-nsp] Ibgp Routes.
>
>
> >
> > thats normal bgp behaviour
> >
> > ibgp does not send full routes unless you configure them to each be route
> > reflector clients.. they will only send the best external routes
> >
> > Steve
> >
> > On Mon, 6 Mar 2006, A. Krijgsman - QCC wrote:
> >
> >> Hi All,
> >>
> >> Does anyone has an explination for this?
> >> - Two routers in VRRP-E.
> >> - IBGP session between them.
> >>
> >> Why don't the IBGP peers sent eachother their entire received
> >> full-transtit
> >> routes?
> >> Does not make sense to me :-(
> >>
> >> Thanks in advance.
> >>
> >> Regards,
> >> Armand
> >>
> >>
> >> Summary:
> >> +++++++++++++++++++++++++++++++++++++++++
> >> RT1:
> >> SSH at ams-gs-rt1>show ip bgp summary
> >> BGP4 Summary
> >> Router ID: 195.144.2.2 Local AS Number : 35829
> >> Confederation Identifier : not configured
> >> Confederation Peers:
> >> Maximum Number of Paths Supported for Load Sharing : 1
> >> Number of Neighbors Configured : 2, UP: 2
> >> Number of Routes Installed : 325556
> >> Number of Routes Advertising to All Neighbors : 118348
> >> Number of Attribute Entries Installed : 61441
> >> Neighbor Address AS# State Time Rt:Accepted Filtered Sent
> >> ToSend
> >> 195.144.2.3 35829 ESTAB 3d 9h59m 143228 0 118347 0
> >> 217.170.9.129 24785 ESTAB 4d12h11m 182327 0 1 0
> >>
> >>
> >> RT2:
> >> SSH at ams-gs-rt2>show ip bgp summary
> >> BGP4 Summary
> >> Router ID: 84.233.188.98 Local AS Number : 35829
> >> Confederation Identifier : not configured
> >> Confederation Peers:
> >> Maximum Number of Paths Supported for Load Sharing : 1
> >> Number of Neighbors Configured : 2, UP: 2
> >> Number of Routes Installed : 297731
> >> Number of Routes Advertising to All Neighbors : 143226
> >> Number of Attribute Entries Installed : 52427
> >> Neighbor Address AS# State Time Rt:Accepted Filtered Sent
> >> ToSend
> >> 84.233.188.97 8928 ESTAB 7d 2h57m 179398 0 1 0
> >> 195.144.2.2 35829 ESTAB 3d 9h58m 118332 0 143225 0
> >>
> >>
> >>
> >>
> >> _______________________________________________
> >> foundry-nsp mailing list
> >> foundry-nsp at puck.nether.net
> >> http://puck.nether.net/mailman/listinfo/foundry-nsp
> >>
> >
> >
>
>
>