03-27-2009 07:33 AM
Hi,
I'm working on the below scenario.
R3 ---------- R1 -------------- R2
150.1.3.3 150.1.1.1 150.1.2.2 -> Lo
LDP is up between R3-R1 and R1-R2
But I do not see any label assigned to R3 & R2 lo IPs on R1 LFIB table.
R1#sh mpls ldp ne
Peer LDP Ident: 150.1.3.3:0; Local LDP Ident 150.1.1.1:0
TCP connection: 150.1.3.3.23155 - 150.1.1.1.646
State: Oper; Msgs sent/rcvd: 33/36; Downstream
Up time: 00:20:56
LDP discovery sources:
Serial2/1, Src IP addr: 150.1.13.3
Addresses bound to peer LDP Ident:
10.1.37.3 10.1.38.3 150.1.13.3 150.1.3.3
Peer LDP Ident: 150.1.2.2:0; Local LDP Ident 150.1.1.1:0
TCP connection: 150.1.2.2.11009 - 150.1.1.1.646
State: Oper; Msgs sent/rcvd: 29/28; Downstream
Up time: 00:16:39
LDP discovery sources:
FastEthernet0/0, Src IP addr: 150.1.12.2
Addresses bound to peer LDP Ident:
150.1.12.2 150.1.2.2 150.1.24.2
R1#sh mpls forward
Local Outgoing Prefix Bytes Label Outgoing Next Hop
Label Label or VC or Tunnel Id Switched interface
16 No Label 150.1.3.3/32 13243 Se2/1 point2point
17 No Label 150.1.2.2/32 0 Fa0/0 150.1.12.2
18 16 150.1.4.4/32 15063 Fa0/0 150.1.12.2
19 Pop Label 150.1.24.0/24 0 Fa0/0 150.1.12.2
R2#sh mpls ldp ne
Peer LDP Ident: 150.1.1.1:0; Local LDP Ident 150.1.2.2:0
TCP connection: 150.1.1.1.646 - 150.1.2.2.11009
State: Oper; Msgs sent/rcvd: 29/29; Downstream
Up time: 00:17:12
LDP discovery sources:
FastEthernet0/0, Src IP addr: 150.1.12.1
Addresses bound to peer LDP Ident:
150.1.12.1 150.1.1.1 150.1.13.1
Peer LDP Ident: 150.1.4.4:0; Local LDP Ident 150.1.2.2:0
TCP connection: 150.1.4.4.23266 - 150.1.2.2.646
State: Oper; Msgs sent/rcvd: 13/16; Downstream
Up time: 00:03:31
LDP discovery sources:
Serial2/0.1, Src IP addr: 150.1.24.4
Addresses bound to peer LDP Ident:
10.1.46.4 10.1.45.4 150.1.4.4 150.1.24.4
R2#sh mpls for
Local Outgoing Prefix Bytes Label Outgoing Next Hop
Label Label or VC or Tunnel Id Switched interface
16 No Label 150.1.4.4/32 13040 Se2/0.1 point2point
17 No Label 150.1.1.1/32 0 Fa0/0 150.1.12.1
18 16 150.1.3.3/32 15316 Fa0/0 150.1.12.1
19 Pop Label 150.1.13.0/24 0 Fa0/0 150.1.12.1
R2#
R3#sh mpls ldp ne
Peer LDP Ident: 150.1.1.1:0; Local LDP Ident 150.1.3.3:0
TCP connection: 150.1.1.1.646 - 150.1.3.3.23155
State: Oper; Msgs sent/rcvd: 36/34; Downstream
Up time: 00:21:36
LDP discovery sources:
Serial1/2, Src IP addr: 150.1.13.1
Addresses bound to peer LDP Ident:
150.1.12.1 150.1.1.1 150.1.13.1
R3#
R3#
R3#
R3#sh mpls fo
Local Outgoing Prefix Bytes tag Outgoing Next Hop
tag tag or VC or Tunnel Id switched interface
16 Untagged 150.1.1.1/32 0 Se1/2 point2point
17 Pop tag 150.1.12.0/24 0 Se1/2 point2point
18 17 150.1.2.2/32 0 Se1/2 point2point
19 18 150.1.4.4/32 0 Se1/2 point2point
20 19 150.1.24.0/24 0 Se1/2 point2point
If LDP neighborship is up, I thought labels will be exchanged automatically for the prefixes. Also CEF is enabled on all devices.
Any thoughts on what could be the problem or how should I troubleshoot this?
Regards
Sourabh
Solved! Go to Solution.
03-27-2009 11:52 AM
Sourabh,
The issue is that you use a /24 subnet mask on the loopback interfaces and that ospf advertises a /32 by default for a loopback interface regardless of what subnet mask is configured. Best practice is to configured a /32 on loopback interfaces. This will definitely solve the issue.
Regards
03-27-2009 08:07 AM
Sourabh,
What platform and IOS release are you using?
Regards
03-27-2009 09:32 AM
R1 -> 7206VXR NPE400 -> 12.2(25)S9
R2 -> 7206VXR NPE400 -> 12.2(25)S9
R3 -> C3640 -> 12.4(13b)
03-27-2009 11:07 AM
Sourabh,
Could you please post the configs.
Regards
03-27-2009 11:27 AM
03-27-2009 11:52 AM
Sourabh,
The issue is that you use a /24 subnet mask on the loopback interfaces and that ospf advertises a /32 by default for a loopback interface regardless of what subnet mask is configured. Best practice is to configured a /32 on loopback interfaces. This will definitely solve the issue.
Regards
03-27-2009 01:57 PM
Hello Harold, Sourabh
as an alternative to have OSPF to advertise loopback addresses with a prefix len less then 32 the same as LDP:
int loop m
ip ospf network point-to-point
However, in production networks /32 loopbacks are the common practice.
Hope to help
Giuseppe
03-27-2009 10:06 PM
Hi Sourabh
Genrally loopbacks in ospf are advertise as stub host because of this they come as /32 as told by hritter. If you want to advertise as same subnet in that case you have to use ip ospf network point-to-point. This is help you to advertise the ospf as mentioned subnet.
regards
shivlu jain
03-28-2009 02:04 AM
Thanks Harold for the help. Problem is resolved now.
07-13-2023 08:56 PM - edited 07-13-2023 08:58 PM
@Harold Ritter your old post solve my problem. my loopback was /24 in that case every time its come with no label on other router now when I configure /32 immediately its reflect other router as PoP Label. Though in both case /24 and /32 the router locally in LIB generate local label as "implicit null".
Now can any one tell me why /24 not showing PoP Label for other router where as both case locally "implicit null" generated in LIB?
(Note:my underlay is ospf and what if underlay is eigrp in this sitution)
07-14-2023 06:47 AM
Glad this old post helped solve the issue @TangoAlfa and thanks for the feedback
07-14-2023 06:49 AM
As Harold has explained, the issue is that you use a /24 subnet mask on the loopback interfaces and that ospf advertises a /32 by default for a loopback interface regardless of what subnet mask is configured.
Other routers did not receive a routing update with prefix/24 mask, but with prefix/32 mask. As they do not have prefix/24 in their routing tables, they ignore the advertised "implicit null" operation for prefix/24 and do not program it in the LFIB. In other words, the routing tables are not indentical on all routers.
This issue is specific for ospf. To advertise a real mask and not /32 you have to configure "ip ospf network point-to-point" under the loopback interface.
03-27-2009 09:08 AM
Hi Sourabh
I have sen the output you are getting no labels on R1 becasue R3 and R2 will advertise its directly conncted interace to R1 as implicit null.
You can test it with loopbacks also.
http://shivlu.blogspot.com/2008/11/what-will-happen-if-loopbacks-are-not.html
regards
shivlu jain
03-27-2009 09:24 AM
Hi Shivlu,
I understand from your blog that this issue will happen in case there is a MP-BGP session created between 2 routers using directly connected IPs.
However, in my case I have not configured MP-BGP yet on any of three devices. First I want to setup LDP between peers so that labels can be exchanged.
I understand that due to PHP, implicit null will be advertised for loopback interfaces, but in that case output on R1 should be 'POP Label' and not 'No Labels'.
Let me know if I'm missing anything.
Regards
Sourabh
03-27-2009 10:10 AM
Shivlu,
An implicit null should be seen as a "Pop Label" not as a "No Label", as the latter will cause all labels to be removed before the packet is sent out the egress interface, which will definitely be an issue if you are carrying a service label.
Regards
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