Re: BGP "no synchronization"

On Mar 30, 11:18 am, Bo...@xxxxxxxxxxxxx wrote:
Hi,

I wonder if anyone can assist with a BGP
understanding issue?

I have a lab setup with BGP and it seems that
I do not understand "bgp synchronisation".

I have 2 AS border routers and a single internal
router with IBGP fully meshed across the three.
There is also another AS clearly but that does
seem to be related to the issue.

The internal router has routes to the bgp next hop
however the routes do not get put in the table
unless I turn off synchronisation.

WITH NO SYNCHRONIZATION
The internal router has

router bgp 1
no synchronization
bgp log-neighbor-changes
neighbor 1.1.1.17 remote-as 1
neighbor 1.1.1.21 remote-as 1
no auto-summary

RH#sh ip rou
1.0.0.0/8 is variably subnetted, 6 subnets, 2 masks
D 1.1.1.0/30 [90/2816] via 1.1.1.17, 00:53:16, FastEthernet2/0
D 1.1.1.0/24 [90/2560] via 1.1.1.21, 00:53:18, FastEthernet3/0
D 1.1.1.4/30 [90/2816] via 1.1.1.21, 00:53:15, FastEthernet3/0
D 1.1.1.12/30 [90/5120] via 1.1.1.21, 00:53:18, FastEthernet3/0
[90/5120] via 1.1.1.17, 00:53:18, FastEthernet2/0
C 1.1.1.16/30 is directly connected, FastEthernet2/0
C 1.1.1.20/30 is directly connected, FastEthernet3/0
2.0.0.0/24 is subnetted, 2 subnets

!! IBGP ROUTES ideal
B 2.2.2.0 [200/0] via 1.1.1.1, 00:03:59
B 2.2.3.0 [200/0] via 1.1.1.5, 00:02:31

RH#sh ip bgp

Network Next Hop Metric LocPrf Weight Path
r i1.1.1.0/24 1.1.1.21 0 100 0 i
r>i 1.1.1.17 0 100 0 i
* i2.2.2.0/24 1.1.1.5 0 100 0 2 i
*>i 1.1.1.1 0 100 0 2 i
*>i2.2.3.0/24 1.1.1.5 0 200 0 2 i
* i 1.1.1.1 0 100 0 2 i
RH#

All OK.

Now turn on synchronisation.

RH(config)#router bgp 1
RH(config-router)#synch

RH#clear ip bgp *

We really have:

router bgp 1
synchronization

RH#sh ip bgp

Network Next Hop Metric LocPrf Weight Path
r>i1.1.1.0/24 1.1.1.17 0 100 0 i
r i 1.1.1.21 0 100 0 i
* i2.2.2.0/24 1.1.1.1 0 100 0 2 i
* i 1.1.1.5 0 100 0 2 i
* i2.2.3.0/24 1.1.1.1 0 100 0 2 i
* i 1.1.1.5 0 200 0 2 i
RH#

RH#sh ip route

1.0.0.0/8 is variably subnetted, 6 subnets, 2 masks
D 1.1.1.0/30 [90/2816] via 1.1.1.17, 00:59:39, FastEthernet2/0
D 1.1.1.0/24 [90/2560] via 1.1.1.21, 00:59:41, FastEthernet3/0
D 1.1.1.4/30 [90/2816] via 1.1.1.21, 00:59:38, FastEthernet3/0
D 1.1.1.12/30 [90/5120] via 1.1.1.21, 00:59:41, FastEthernet3/0
[90/5120] via 1.1.1.17, 00:59:41, FastEthernet2/0
C 1.1.1.16/30 is directly connected, FastEthernet2/0
C 1.1.1.20/30 is directly connected, FastEthernet3/0
RH#

Since I have EIGRP routes to the BGP next hop
i.e. 1.1.1.1 - 1.1.1.0/30 [90/2816] via 1.1.1.17
1.1.1.5 - 1.1.1.4/30 [90/2816] via 1.1.1.21

I would have though that I was Synchronised and would not
need NO SYCHRONIZATION in the BGP config??

Clearly I am mistaken.

Anyone care to explian?

BSCI tomorrow - don't suppose I will need this but
I have gone this far and would like to understand it
while the iron is hot as it were.


It is NOT the next hop that has to be learned via the IGP, it is the
BGP route/prefix ITSELF that must be learned via the IGP.

So in your example 2.2.2.0/24 and 2.2.3.0/24 would have to be
learned via EIGRP before an IBGP router should install the route

And to test BGP should not be runninging on the Internal router just
the IGP.

So remove BGP from you internal route advertise 2.2.2.0/24 on one of
the external routers and see if your other external router is then
willing to install with


The BGP synronization feature was meant to be used for the scenario
where you are providing transit but not all routers in the path are
running BGP. in other words the traffic would be dropped at the
internal router not running BGP because obviously there is no BGP
route and there is no IGP route EITHER.

This feature is never used in an ISP environment as far as I know. On
IOS XR for CRS, no sync is the default unlike IOS where sync is still
the default
.

Relevant Pages

  • Re: BGP
    ... This should work, except the null route ... source router, ... bgp log-neighbor-changes ... network 172.172.172.168 mask 255.255.255.255 ...
    (comp.dcom.sys.cisco)
  • Re: Load-balancing across four T1s on 2 routers
    ... router will receive 2/3 of the traffic and the other will receiving 1/3 ... EIGRP supports UNequal-cost load-balancing in that manner, ... In you case BGP scenario will be really easy. ... switch will then see 2 UN-equal cost default routes in its routing ...
    (comp.dcom.sys.cisco)
  • Re: VPN Client Routing Issues
    ... With regards to static redistribution into BGP, ... Is the route 192.168.237.0/24 in the local CE router BGP routing table ... (the onsite MPLS router)? ... Is the route being advertised via BGP to the upstream MPLS PE router? ...
    (comp.dcom.sys.cisco)
  • Re: BGP "no synchronization"
    ... I do not understand "bgp synchronisation". ... router with IBGP fully meshed across the three. ... learned via EIGRP before an IBGP router should install the route ...
    (comp.dcom.sys.cisco)
  • Re: Load-balancing across four T1s on 2 routers
    ... UNequal-cost load-balancing? ... router will receive 2/3 of the traffic and the other will receiving 1/3 ... just add couple more command to your BGP config ... EIGRP between you and ISP? ...
    (comp.dcom.sys.cisco)