06-04-2009 05:53 AM - edited 03-04-2019 04:59 AM
Hi,
Recently we started to route all our outgoing National call though our carriers network to get cheaper calls. Since we rolled this out we have noticed that all these calls are routing across two sites before it leaves our network.
I have attached a screen shot of a diagram of the main part of our network.
We have a Cisco Call Manager 4.1.3(SR7) and there is a gateway created for these calls which I have also attached.
When I run a traceroute on the HQ Voice Gateway, the path to 20.40.60.80 router to the Router on HQ side of the 10MB leased line to the router on the Call Centre side of the leased line and then routes across the Call Centre to the Call Centre 10MB link to the Carrier Network. It does the same for the HA site.
I would like the traffic designated for the Carried network to leave the HA and HQ site without going through our Call Centre.
Where should I inject a route to make these call exit the network locally.
Thanks a million
Anthony.
06-04-2009 12:29 PM
Hello Anthony,
I would suggest you to provide the following information:
on HQ Voice Gateway
sh ip protocol summary
sh ip route 20.40.60.80
trace 20.40.60.80
This in order to get better help.
Hope to help
Giuseppe
06-04-2009 11:37 PM
Hi Giuseppe, please see info that you requested,
HQ_VG#sh ip protocol sum
Index Process Name
0 connected
1 static
2 ospf 1
HQ_VG#sh ip route 20.40.60.80
Routing entry for 20.40.60.80/24
Known via "ospf 1", distance 110, metric 100
Tag 5466, type extern 2, forward metric 11
Last update from 1.2.3.4 on GigabitEthernet0/0, 5w1d ago
Routing Descriptor Blocks:
* 1.2.3.4, from 10.1.1.1, 5w1d ago, via GigabitEthernet0/0
Route metric is 100, traffic share count is 1
Route tag 5466
HQ_VG#trace 86.43.21.25
Type escape sequence to abort.
Tracing the route to 20.40.60.80
1 1.2.3.4 4 msec 0 msec 0 msec
2 10.1.1.1 4 msec 4 msec 12 msec
3 10.10.1.1 12 msec 8 msec 8 msec
4 30.60.90.120 8 msec 12 msec 12 msec
5 * * *
6 * * *
7 * * *
8 * * *
9 * * *
10 * * *
11 * * *
12 * * *
13 * * *
14 * * *
15 * * *
16 * * *
17 * * *
18 * * *
19 * * *
20 * * *
21 * * *
22 * * *
23 * * *
24 * * *
25 * * *
26 * * *
27 * * *
28 * * *
29 * * *
30 * * *
The 30.60.90.120 is the gateway out of our Call Centre network to the Carrier's network.
Thanks a million for your help
Anthony.
06-05-2009 12:05 AM
Hello Anthony,
the prefix 20.40.60.80 is an external route of type 2 (O E2 in routing table) learned by ospf process 1.
O E2 means the external seed metric 100 is not added to the internal path metric
The OSPF forwarding metric is 11.
The router will use the shortest internal path. (fwd metric 11)
The point is to see what would be the cost on the desired path.
To see this I would suggest you to provide the following:
sh ip ospf database external 20.40.60.80
sh ip ospf interface type x/y
where typex/y is the desired outgoing interface on VGW1
Hope to help
Giuseppe
06-05-2009 02:10 AM
Hi Giuseppe.
Please see the info you requested.
HQ_VG#sh ip ospf database external 20.40.60.80
OSPF Router with ID (192.168.0.1) (Process ID 1).
The Carriers network is attached to our HQ_WR (WAN Router) and the calls currently routes via these hops:
HQ_VG
HQ Call Centre on HQ's side
HQ Call Centre on Call Centre Side
Call Centre WR
Carrier Network
I would like it to go via
HQ_VG
HQ_WR
Carrier Network (This interface is a ATM2/0)
When looking into this, i'm getting confused. As far as I know I want it to do out the Gi0/0 on the VG to the HQ_WR and from the HQ_WR out the ATM interface and onto the Carrier's network
Am I going in the right direction
Thanks a million for your help
Anthony.
06-05-2009 06:21 AM
06-05-2009 09:11 AM
Hello Anthony,
I see that also iBGP is involved in this scenario but it is not running on the HQ VGW because sh ip prot sum says only OSPF is running and static routes.
To further investigate we need to find where the ASBR node (the router that originates the 20.40.60.80 route) is placed in your network diagram.
I may be wrong but it should be a node in the call centre.
In addition to this in order to find a fix it is useful to check if the HQ 3825 that is running iBGP is also running OSPF.
To have the HQ VGW to use HQ 3825 to send traffic destined to 20.40.60.80 this node HQ 3825 the following conditions should be met:
HQ 3825 should
1) run OSPF and to take part in same OSPF domain
2)to have learned the required route by any means it is likely it learns it from iBGP. you can check this on HQ router by using sh ip route 20.40.60.80 and trace 20.40.60.80 verify it sees
3) HQ 3825 has to publish in OSPF its knowledge of prefix 20.40.60.80 with a more attractive metric ( a metric lower then 100 or better to use type O E1 that is preferred).
Note:
it is not acceptable for you to use the two GE links between HQ and HA ?
this could even easier if the prefix 20.40.60.80 is advertised by HA multilayer switches over the GE links in this case the prefix would become an O IA or O route (if intra area) and would be again preferred to current O E2 route.
Hope to help
Giuseppe
06-08-2009 01:22 AM
Hi Giuseppe, Hope you had a good weekend.
I ran the following from the HQ Wan Router and it does learn the router from BGP and if the HQWR need to go to 20.40.60.80 it goes out the 10MB BIP link to the Carrier.
HQWR#sh ip proto sum
Index Process Name
0 connected
1 static
2 ospf 1
3 bgp 65500
HQWR#
HQWR#
HQWR#sh ip route 20.4.60.80
Routing entry for 20.4.60.80/24
Known via "bgp 65500", distance 20, metric 0
Tag 5466, type external
Redistributing via ospf 1
Last update from 10.20.30.40 5w4d ago
Routing Descriptor Blocks:
* 10.20.30.40, from 10.20.30.40, 5w4d ago
Route metric is 0, traffic share count is 1
AS Hops 1
Route tag 5466
HQWR#trace 20.40.60.80
Type escape sequence to abort.
Tracing the route to 20.40.60.80
1 10.20.30.40 0 msec 8 msec 4 msec
I hope all this helps. If you have4 any questions, please let me know.
Thanks
Anthony.
06-08-2009 11:07 PM
Hello Anthony,
sorry for late reply
from the following:
HQWR#sh ip route 20.4.60.80
Routing entry for 20.4.60.80/24
Known via "bgp 65500", distance 20, metric 0
Tag 5466, type external
Redistributing via ospf 1
Last update from 10.20.30.40 5w4d ago
Notice that the route is learned from BGP but one line is missing:
advertised by OSPF 1
my guess is that the route is not advertised in OSPF.
you can check this on HQWR with
sh ip ospf database external 20.40.60.80
if so you need to modify the route filter that decides what is redistributed from BGP into OSPF and also there is a special command in BGP that instructs ot redistribute iBGP routes.
router bgp xx
bgp redistribute-internal
see
http://www.cisco.com/en/US/docs/ios/iproute/command/reference/irp_bgp1.html#wp1014183
Hope to help
Giuseppe
06-08-2009 11:42 PM
Hi Giuseppe,
Currently, BGP is setup as follows:
router bgp 40000
no synchronization
bgp log-neighbor-changes
network 15.30.45.60 mask 255.255.255.252
network LINK TO CARRIER BIP mask 255.255.255.252
redistribute ospf 1 route-map local-networks
neighbor HA WAN RTR remote-as 40000
neighbor HA WAN RTR next-hop-self
neighbor HA WAN RTR distribute-list 4 out
neighbor CARRIER BIP NODE INTO HQ remote-as 5466
neighbor CARRIER BIP NODE INTO HQ send-community
neighbor CARRIER BIP NODE INTO HQ distribute-list 1 in
neighbor CARRIER BIP NODE INTO HQ distribute-list 8 out
neighbor CARRIER BIP NODE INTO HQ route-map increase-odd-localpref in
neighbor CARRIER BIP NODE INTO HQ route-map odd-community out
distance bgp 20 20 25
distance 115 CARRIER BIP NODE INTO HQ 0.0.0.0 CallCentre-prefixes
distance 115 HA WAN RTR 0.0.0.0 CallCentre-prefixes
no auto-summary
Just to confirm what I need to do:
router bgp 40000
bgp redistribute-internal
and do some test calls and see where they route out.
Thanks a million for your help
Anthony.
Discover and save your favorite ideas. Come back to expert answers, step-by-step guides, recent topics, and more.
New here? Get started with these tips. How to use Community New member guide