From: Braychuck Vitaliy (Vitaliy.Braychuck@incom.ua)
Date: Fri Jan 23 2009 - 07:40:09 ARST
Could you try ping 4.4.4.4 source <interface with ip address 141.41.67.6>
-----Original Message-----
From: nobody@groupstudy.com [mailto:nobody@groupstudy.com] On Behalf Of joe_astorino@comcast.net
Sent: Friday, January 23, 2009 11:23 AM
To: joe_astorino@comcast.net; ccielab@groupstudy.com
Subject: Re: Frame-Relay Encapsulation Failed -- Driving me mad
Sorry I meant if I add a frame map to 4.4.4.4 it works
Sent from my Verizon Wireless BlackBerry
-----Original Message-----
From: joe_astorino@comcast.net
Date: Fri, 23 Jan 2009 08:01:34
To: <ccielab@groupstudy.com>
Subject: Frame-Relay Encapsulation Failed -- Driving me mad
Hello everybody. I am hoping somebody might be able to explain to me an issue I am having regarding frame-relay that is driving me mad in the practice lab. Here is the scenario:
R2, R5, R6 are in a frame-relay hub/spoke topology whereby R2 is the hub. DLCIs are X0Y where X is the source router number and Y is the destination router number. Additionally, R5 has a PPPoFR link to R4. The R2,R5,R6 cloud is the 141.41.26.0/24 network. The R5/R4 P2P is the 141.141.45.0/24 network.
I have full ip reachability, all my routing is working fine and I run into this IOS services task.
R4 has an ethernet interface on the 141.141.200.0/24 subnet. R6, which is on the other end of the frame-relay cloud from R4 has an ethernet interface on the 141.41.67.0/24 subnet. The task requires that you configure R4 to be a DHCP server to hand out addresses on the 141.41.67.0/24 subnet. In other words, you need a helper address on R6's ethernet. In order to verify your configuration, you need to bring up R8's ethernet interface and set it for dhcp.
Here is where it gets weird. On R6 I have the following frame mappings as you would expect of a spoke: A static mapping to the hub, the other spoke, and itself.
R6(config-if)#do sh frame map
Serial2/0 (up): ip 141.41.26.2 dlci 602(0x25A,0x94A0), static,
broadcast,
CISCO, status defined, active
Serial2/0 (up): ip 141.41.26.5 dlci 602(0x25A,0x94A0), static,
CISCO, status defined, active
Serial2/0 (up): ip 141.41.26.6 dlci 602(0x25A,0x94A0), static,
CISCO, status defined, active
After I have everything setup, R8 is not pulling a DHCP address, so I turn on some debugging and see this:
..Jan 23 07:52:32.443: IP: s=0.0.0.0 (Ethernet0/0), d=255.255.255.255, len 604, rcvd 2
.Jan 23 07:52:32.443: UDP src=68, dst=67
.Jan 23 07:52:32.443: IP: s=141.41.67.6 (local), d=4.4.4.4 (Serial2/0), len 604, sending
.Jan 23 07:52:32.443: UDP src=67, dst=67
.Jan 23 07:52:32.447: IP: s=141.41.67.6 (local), d=4.4.4.4 (Serial2/0), len 604, encapsulation failed
.Jan 23 07:52:32.447: UDP src=67, dst=67
So, I see the DHCP broadcast come in from R8, thats good. Then I see R6 attempt to unicast to R4 because it has ip helper-address 4.4.4.4 configured (R4's loopback) so thats good ....... and then boom encapsulation failed. First off I don't understand why this is happening. I have a route to 4.4.4.4, which R6 learnes via ospf from R5 from across the frame. I would think that like ANY other route it has that it would first lookup 4.4.4.4 in the routing table, see that the next hop is 141.41.26.5, then send the packet source 141.41.67.6 , destination 4.4.4.4 with a DLCI of 602 as it is destined for R5. Here is the route entry on R6.
Routing entry for 4.4.4.4/32
Known via "ospf 1", distance 110, metric 20, type extern 2, forward metric 64
Redistributing via eigrp 679
Advertised by eigrp 679 metric 1 1 1 1 1
Last update from 141.41.26.5 on Serial2/0, 01:37:43 ago
Routing Descriptor Blocks:
* 141.41.26.5, from 5.5.5.5, 01:37:43 ago, via Serial2/0
Route metric is 20, traffic share count is 1
No, here is the really insane part! If I change absolutely NOTHING and ping 4.4.4.4 from R6 it works just fine. If I ping ANY other route in my lab network, it works just fine. I am stumped, hoping for some help!
Here is the first packet sent, and first received: So why is it that if I send an icmp packet it works, but if I send a UDP packet to the exact same ip address it fails encapsulation? FYI there are no ACLs or filtering involved. Also if I do a frame map to 141.141.45.4 out DLCI 602 it works. I just don't get why I would need that at all, since as I said above I think it should recursively look up the route, and send it out the proper DLCI to the next hop address.
R6(config-if)#do ping 4.4.4.4
Type escape sequence to abort.
Sending 5, 100-byte ICMP Echos to 4.4.4.4, timeout is 2 seconds:
!!
.Jan 23 07:57:37.352: IP: tableid=0, s=141.41.26.6 (local), d=4.4.4.4 (Serial2/0), routed via FIB
.Jan 23 07:57:37.352: IP: s=141.41.26.6 (local), d=4.4.4.4 (Serial2/0), len 100, sending
.Jan 23 07:57:37.352: ICMP type=8, code=0
.Jan 23 07:57:37.548: IP: tableid=0, s=4.4.4.4 (Serial2/0), d=141.41.26.6 (Serial2/0), routed via RIB
.Jan 23 07:57:37.548: IP: s=4.4.4.4 (Serial2/0), d=141.41.26.6 (Serial2/0), len 100, rcvd 3
.Jan 23 07:57:37.548: ICMP type=0, code=0
Thanks for ANY help on this one!
- Joe
Blogs and organic groups at http://www.ccie.net
This archive was generated by hypermail 2.1.4 : Sun Mar 01 2009 - 09:43:39 ARST