Troubleshooting Frame Relay 1

[snickm]

Hi.

I am having trouble with a frame relay connection that was recently increased from 192k to 1Mb. Communication in one direction is fine but in the other direction a ping will drop every second packet. Below is the output from a Windows 2000 server running a traceroute to the Cisco router at the other end of the Frame Relay connection:

C:\WINNT40\Profiles\Administrator>tracert 10.1.1.254

Tracing route to 10.1.1.254 over a maximum of 30 hops

1 <10 ms <10 ms <10 ms 10.1.2.254
2 * 31 ms * 10.1.1.254
3 31 ms * * 10.1.1.254
4 * 31 ms * 10.1.1.254
5 * 31 ms * 10.1.1.254
6 * 31 ms * 10.1.1.254
7 * 31 ms * 10.1.1.254
8 * 31 ms * 10.1.1.254
9 * 31 ms * 10.1.1.254
10 * 31 ms * 10.1.1.254
11 * 31 ms * 10.1.1.254
12 * 31 ms * 10.1.1.254
13 * 31 ms * 10.1.1.254
14 * 31 ms * 10.1.1.254
15 * 31 ms * 10.1.1.254
16 * 31 ms * 10.1.1.254
17 * 31 ms * 10.1.1.254
18 * 31 ms * 10.1.1.254
19 * 31 ms * 10.1.1.254
20 * 31 ms * 10.1.1.254
21 * 31 ms * 10.1.1.254
22 * 31 ms * 10.1.1.254
23 * 31 ms * 10.1.1.254
24 * 31 ms * 10.1.1.254
25 * 31 ms * 10.1.1.254
26 * 31 ms * 10.1.1.254
27 * 31 ms * 10.1.1.254
28 * 31 ms * 10.1.1.254
29 * 31 ms * 10.1.1.254
30 * 31 ms * 10.1.1.254

Trace complete.

Kinda strange that the router is accepting the packet then boncing it around inside itself. Next is the routing table on the Cisco router:

router#sh ip route

Gateway of last resort is 10.1.1.250 to network 0.0.0.0

S 192.43.198.0/24 is directly connected, Serial3/0.17
192.168.4.0/30 is subnetted, 2 subnets
C 192.168.4.4 is directly connected, Serial3/0.17
C 192.168.4.0 is directly connected, Serial3/0.18
10.0.0.0/8 is variably subnetted, 6 subnets, 2 masks
C 10.1.3.0/24 is directly connected, Loopback0
S 10.1.2.0/24 is directly connected, Dialer1
is directly connected, Serial3/0.1
C 10.1.1.0/24 is directly connected, Ethernet0/0
S 10.7.2.0/24 is directly connected, Serial3/0.18
C 10.1.1.91/32 is directly connected, Async35
C 10.128.1.0/24 is directly connected, Dialer1
S 192.168.7.0/24 [1/0] via 10.1.1.250
192.168.3.0/30 is subnetted, 1 subnets
C 192.168.3.0 is directly connected, Serial3/0.1
S* 0.0.0.0/0 [1/0] via 10.1.1.250

I have the Telecoms company checking into the Frame Relay connection for me, can anyone shed some light on this problem? Have you seen it or something similar before?

 

[tmt7734]

Did yuo hear back from telco yet? Maybe the PVC has a phantom DLCI pointing to itself. What do you see if you do a sho fra pvc?

 

[dearingkr]

One possible issue I see is that you have ths same subnet on multiple interfaces.

192.168.4.x on 2 different interfaces. With a 30-bit subnet mask?

Also 10.1.1.x/24 is on 2 different interfaces.

Seems like this would give you all kinds of strange routing problems.

MCSE CCNA CCDA

 

[pmesjar]

Dearingkr, 192.168.4.x is fine subnetted in his way as long as he is using point-to-point subinterfaces and for that you need 30bit subnet mask.

snickm, you did not provide more info, like source IP address you are doing traceroute from. Maybe your default route points into wrong address, because you have the whole 10.1.1.x/24 subnet assigned to Ethernet0/0 interface, but your default route points to 10.1.1.250, which is from this subnet. Which device your default route points to anyway? If the default route is configured right, it must be some device connected to Ethernet0/0. How about sending a full config of the router? Also "sh frame pvc" and "sh frame map" can be helpful.

Peter Mesjar
CCNP, A+ certified
pmesjar@centrum.sk

"The only true wisdom is in knowing you know nothing.

 

[snickm]

Thanks for the replies guys. Sorry 'bout the delay- don't usually go to work on a weekend. Anyways, to answer your questions:

Source IP of the traceroute is 10.1.2.1

Default route 10.1.1.250 is our firewall, this router sits inside the firewall.

Router config:

router#sh run
Building configuration...

Current configuration:
!
version 12.0
service timestamps debug uptime
service timestamps log uptime
service password-encryption
service udp-small-servers
service tcp-small-servers
!
hostname
!
logging console informational
enable secret
enable password
!
username
ip subnet-zero
no ip domain-lookup
ip name-server 10.1.1.5
isdn switch-type primary-net5
!
!
controller E1 2/0
pri-group timeslots 1-10,16
!
!
interface Loopback0
description Loopback for Dial up users
ip address 10.1.3.254 255.255.255.0
no ip directed-broadcast
!
interface Ethernet0/0
description Connected to LAN 10.1.1.0
ip address 10.1.1.254 255.255.255.0
no ip directed-broadcast
!
interface Serial2/0:15
description Master Dial In Controller
no ip address
no ip directed-broadcast
encapsulation ppp
ip mroute-cache
dialer pool-member 1
isdn switch-type primary-net5
isdn incoming-voice modem
ppp authentication pap chap
!
interface Ethernet3/0
no ip address
no ip directed-broadcast
shutdown
!
interface Serial3/0
description Frame Relay Service
no ip address
no ip directed-broadcast
encapsulation frame-relay
logging event subif-link-status
logging event dlci-status-change
frame-relay lmi-type ansi
!
interface Serial3/0.1 point-to-point
description Frame Relay from Remote Site 1
bandwidth 1024
ip address 192.168.3.1 255.255.255.252
no ip directed-broadcast
no arp frame-relay
frame-relay interface-dlci 16
!
interface Serial3/0.17 point-to-point
description Frame Relay from Remote Site 3
bandwidth 256
ip address 192.168.4.5 255.255.255.252
no ip directed-broadcast
no arp frame-relay
frame-relay interface-dlci 17
!
interface Serial3/0.18 point-to-point
description Frame Relay from Remote Site 2
bandwidth 128
ip address 192.168.4.1 255.255.255.252
no ip directed-broadcast
no arp frame-relay
frame-relay interface-dlci 18
!
interface Group-Async1
ip unnumbered Ethernet0/0
no ip directed-broadcast
encapsulation ppp
dialer in-band
dialer idle-timeout 600
dialer-group 1
async mode interactive
peer default ip address dhcp
no cdp enable
ppp authentication pap
group-range 33 44
!
interface Dialer1
description In-dial from Remote Site 1
ip address 10.128.1.1 255.255.255.0
no ip directed-broadcast
encapsulation ppp
dialer remote-name
dialer idle-timeout 2000000
dialer pool 1
dialer-group 1
priority-group 1
ppp authentication chap
!
ip local pool dialup 10.1.3.1 10.1.3.50
ip classless
ip route 0.0.0.0 0.0.0.0 10.1.1.250
ip route 10.1.2.0 255.255.255.0 Dialer1
ip route 10.1.2.0 255.255.255.0 Serial3/0.1
ip route 10.7.2.0 255.255.255.0 Serial3/0.18
ip route 192.43.198.0 255.255.255.0 Serial3/0.17
ip route 192.168.7.0 255.255.255.0 10.1.1.250
!
priority-list 1 protocol ip high tcp telnet
priority-list 1 default low
dialer-list 1 protocol ip permit
!
line con 0
transport input none
line 33 44
autoselect during-login
autoselect ppp
login local
modem Dialin
modem autoconfigure type mica
transport input all
line aux 0
line vty 0 4
password
login
!
end

The ISDN line for remote site 1 is shutdown at the other end and only activated if the Frame Relay goes down.

hfcrt2wc#sh frame pvc

PVC Statistics for interface Serial3/0 (Frame Relay DTE)

DLCI = 16, DLCI USAGE = LOCAL, PVC STATUS = ACTIVE, INTERFACE = Serial3/0.1

input pkts 44915954 output pkts 51052732 in bytes 4152021273
out bytes 2556205300 dropped pkts 8 in FECN pkts 4346
in BECN pkts 7 out FECN pkts 0 out BECN pkts 0
in DE pkts 8345135 out DE pkts 0
out bcast pkts 141698 out bcast bytes 43567755
pvc create time 13w6d, last time pvc status changed 5d17h

DLCI = 17, DLCI USAGE = LOCAL, PVC STATUS = ACTIVE, INTERFACE = Serial3/0.17

input pkts 501976 output pkts 480820 in bytes 154156439
out bytes 224008087 dropped pkts 0 in FECN pkts 605
in BECN pkts 0 out FECN pkts 0 out BECN pkts 0
in DE pkts 56954 out DE pkts 0
out bcast pkts 8104 out bcast bytes 2571630
pvc create time 5d18h, last time pvc status changed 5d13h

DLCI = 18, DLCI USAGE = LOCAL, PVC STATUS = ACTIVE, INTERFACE = Serial3/0.18

input pkts 1653522 output pkts 1545805 in bytes 297680998
out bytes 778467105 dropped pkts 0 in FECN pkts 1398
in BECN pkts 0 out FECN pkts 0 out BECN pkts 0
in DE pkts 116789 out DE pkts 0
out bcast pkts 15527 out bcast bytes 4908445
pvc create time 1w3d, last time pvc status changed 18:13:27


Serial3/0.17 (up): point-to-point dlci, dlci 17(0x11,0x410), broadcast
status defined, active
Serial3/0.1 (up): point-to-point dlci, dlci 16(0x10,0x400), broadcast
status defined, active
Serial3/0.18 (up): point-to-point dlci, dlci 18(0x12,0x420), broadcast
status defined, active

Thanks again for your help.

 

[pmesjar]

What I think is that your router will try to load-balance returning packets. Looking at your routing table:

S 10.1.2.0/24 is directly connected, Dialer1
is directly connected, Serial3/0.1

Path to 10.1.2.0/24, from where you source your traffic, is bound to Dialer1 and Serial3/0.1 interfaces and they are made equal. I guess every other packet is dropped because router is not that fast in establishing a connection via Dialer1 interface, while it is fast enough to send it via frame relay circuit. I think you should configure route pointing via Dialer1 interface with higher administrative distance, like:

ip route 10.1.2.0 255.255.255.0 Dialer1 10
ip route 10.1.2.0 255.255.255.0 Serial3/0.1

Hope this helps

And one more thing, before other users start bombarding me with stuff like "per-packet and per-destination load balancing" - I know about these but so far can't think about any other solution to snickm's problem...

Peter Mesjar
CCNP, A+ certified
pmesjar@centrum.sk

"The only true wisdom is in knowing you know nothing.

 

[snickm]

Thanks pmesjar.

As mentioned by my previous post, the ISDN line (connected via Dialer1) is shut down from the other end so no packets can be returned via it. When I removed that route last night it solved our problem immediately.

I'll just have to modify the procedure for bringing the ISDN up to include adding that route should the frame relay ever go down.

Again thanks.

 

[narbik]

Your addressing is all messed up 10.1.1.254 is a network within your ethernet interface.

 

[pmesjar]

snickm, modifying admninistrative distance as I showed you is one way.

narbik, why do you think his addressing is messed up? I think his addressing scheme is does not look nice but so far I can't seem to spot any problems due to messed up addressing. Where do you see the problem?

Peter Mesjar
CCNP, A+ certified
pmesjar@centrum.sk

"The only true wisdom is in knowing you know nothing.

 

[snickm]

I'm curious why you say the addressing "does not look nice"? I haven't done a lot of work with Cisco and IP addressing but I can't think of any other ways to do it. We are using 192.168.?.? inside the serial connections and 10.?.?.? for the ethernet networks. Ignore the 192.43.198.0 network for now as it will be changed to 10.7.3.0 on Friday. The only exception to this rule is the DMZ.

I'd be interested in any suggestions that may improve things.

 

[pmesjar]

You are mixing IP addresses too much. From the routing table I can see you use:

192.168.3.0/30 <-- why? why not rather eg. 192.168.4.8/30?
192.168.4.0/30
192.168.4.4/30
192.168.7.0/24 <-- why? why not rather eg. 192.168.2.0/24?
and why not start from 192.168.0.0?
10.1.3.0/24
10.1.2.0/24
10.1.1.0/24
10.7.2.0/24 <-- why? why not rather eg. 10.1.4.0/24?
10.128.1.0/24 <-- why not rather eg. 10.1.5.0/24

In my opinion addressing should be made more continuous, so you can properly summarize for example. I am not saying this is bad scheme, what I mean under the word "does not look nice" is it surely can be done better. But this also depends on your company's policy and whether those numbers in IP subnets mean something more to you than just IP subnet.


Peter Mesjar
CCNP, A+ certified
pmesjar@centrum.sk

"The only true wisdom is in knowing you know nothing.