Cisco to Cisco T1 issues 1

[jmlarsen]

Given the following setup:

ISP-->RouterA-->T1<--RouterB<--Office

RouterA = Cisco3640
RouterB = Cisco2524

From anywhere on the internet I can ping and telnet to either Router A or B. From an outside source I get no packet loss when pinging RouterA, and about 6% packet loss when pinging RouterB (over the T1).

Now here is the problem. From any host at the 'Office' location, I can ping, traceroute and do dns queries to outside hosts just fine. But as soon as I try to do anything else, I get massive packet loss. Telnet is possible, but very crippled (speed wise), web browsing doesn't get past the name resolution.

I can't figure out what the problem is. Here is the config from the 2524 (RouterB):

Current configuration:
!
version 12.0
service timestamps debug uptime
service timestamps log uptime
service password-encryption
!
hostname RouterB
!
boot system flash c2500-io-l.120-5.T.bin
enable secret 5 $1$IoWc$TYPNZtv6zxhTjmQxS/sg4/
enable password 7 110B1817
!
!
!
ip subnet-zero
no ip finger
no ip domain-lookup
!
!
!
process-max-time 200
!
interface Loopback0
no ip address
no ip directed-broadcast
!
interface Ethernet0
description connected to EthernetLAN
ip address 44.214.131.129 255.255.255.240
no ip directed-broadcast
no keepalive
no cdp enable
!
interface Serial0
bandwidth 1536
no ip address
no ip directed-broadcast
encapsulation frame-relay IETF
no keepalive
no fair-queue
service-module t1 remote-alarm-enable
!
interface Serial0.1 point-to-point
description connected to Cisco3640
bandwidth 1536
ip unnumbered Ethernet0
no ip directed-broadcast
no cdp enable
frame-relay interface-dlci 100 IETF
!
interface Serial1
no ip address
no ip directed-broadcast
no ip mroute-cache
shutdown
no cdp enable
!
router eigrp 200
network 64.0.0.0
no auto-summary
!
ip classless
ip route 0.0.0.0 0.0.0.0 Serial0.1
no ip http server
!
no cdp run
snmp-server engineID local 00000009020000E01E689E19
snmp-server community public RO
!
line con 0
transport input none
line aux 0
transport input all
line vty 0 4
exec-timeout 0 0
password 7 15140403
login
!
end



Any suggestions would be appreciated.

-jml

 

[davidarndt]

I had a similar problem and it turned out that the telco wasn't providing clocking signals on the T1 line. I had to turn on the internal clock on one of the CSU's.

I'm not in front of a Cisco router but the syntax was something like:

service-module T1 clock source internal

After enabling the clock, the routers sync'ed up properly and packet loss went to near zero.

 

[jeter]

A point to point T1 does not provide any clocking, these parameters need to be setup on the routers. One needs to be internal and the other needs to be external. I am also wondering that if this is a PPP then why is there Frame-relay programming involved?
Jeter@LasVegas.com
J.Fisher CCNA

 

[prodsup]

I would like to confirm the answers you have had. This is similar to UK unstructured g703 megastreams, which require you to provide the clocking.

 

[jmlarsen]

You are right, frame-relay is not necessary. I have removed all frame-relay lines. Now it is:

eth0->ser0(ip unnum)-->T1<--ser0/0(ip unnum)<-fastEth0/0

HDLC encap.

3640 is providing clock source
2524 is clock source is line

Still getting same behavior. I can ping the routers from each other (with varying degree of success... 20-80% loss). I can perform ping, traceroute, and DNS lookups from behind the 2524 (over the T1). But nothing else.

One thing to note, I do get Path Code Violations from a 'sh service-module'. Wouldn't that indicate a possible problem in the line?

I'm still frustrated with this. Any more suggestions??

 

[jeter]

Lets do this set your serial interface up and create a sub interface. So you will have a serial0 and a serial0.1 point-to-point lets put the ip address on 3640 as 192.168.1.1 255.255.255.252 and on the 2524 192.168.1.2 . So we have a interface and a sub-interface. Add a static route on the 3640 ( 0.0.0.0 0.0.0.0 192.168.1.2 ) and on the 2524
(0.0.0.0 0.0.0.0 192.168.1.1 ) make sure that ip classless is active. Im concerned about the ethernet ip also lets change them to a 10.1.1.1 and a 10.1.2.1 with a 255.255.255.0 submask. This should get you across and be able to ping from pc to pc after you set them up.

Note if this is a full T1 then you will not have to add this
statement service-module T1 timeslot 1-24 however if it is differnt then add 1-2=128 1-4=256 1-8=512 etc etc
Jeter@LasVegas.com
J.Fisher CCNA

 

[jmlarsen]

Ok, I've got the circuit up again, without frame-relay. Just point to point T1. I am seeing the same behavior. I can ping any host, I can traceroute to any outside host, and I can perform DNS queries/transfers. I can't do anything else though. No HTTP, FTP, SSH, etc. I get varying packet loss from pings (ranging from 20% loss to 90% loss).

The only odd thing I can see is this: (from 'sh service-module)


-2524-

Total Data (last 73 15 minute intervals):
678 Line Code Violations, 1313 Path Code Violations
1 Slip Secs, 771 Fr Loss Secs, 5 Line Err Secs, 0 Degraded Mins
1275 Errored Secs, 40 Bursty Err Secs, 15 Severely Err Secs, 755 Unavail Secs


-3640-

Total Data (last 73 15 minute intervals):
3 Line Code Violations, 536 Path Code Violations
36 Slip Secs, 823 Fr Loss Secs, 2 Line Err Secs, 6 Degraded Mins
15 Errored Secs, 3 Bursty Err Secs, 11 Severely Err Secs, 815 Unavail Secs



It doesn't appear to be a routing issue, since I can ping and traceroute to any outside host. Bell claims the line is fine, they said they were able to loop the CSU and smart jack on the 2524 end, but could not loop the CSU on the 3640 end. What does that indicate? I've even tried replacing the T1/CSU-DSU module in the 3640 on a suggestion of bad hardware.

Any ideas??

Thanks.

 

[davidarndt]

I agree with you that it looks like a layer 1 or 2 problem. Have you tried pinging with a large packet size? I'm guessing that would start causing failures. Might be worth trying just to confirm that it's not a network layer problem.

If you watch the CD light on your CSU, do you see it going off when the errors are occuring? That would definitely implicate line or clocking issues.

 

[jmlarsen]

Found out what the problem was.

The data center where the 3640 is located had a bad cross connection from Verizon.

They re-terminated and everything is fine.

Thanks for the help and suggestions.

 

[cpace]

I am having a similar problem with packet loss. Do I have to do any clocking with frame-relay ? I always assumed this was done at the switch and I have never done this with any of my other frame lines. I have full 1.54 with 128 CIR what syntex would I use ? Thanks in advance.

 

[svermill]

First of all, telco can and often does provide clock on a p-t-p T-1. This is done by routing the T-1 through a Digital Access and Cross-connect System (DACS). I routinely specify that telco provide clock in this way if my customer doesn't have a good source themselves. For instance, the internal oscillator in a router is a very low-grade clock (stratum 3 or less). The telco DACS will be tied back to a precision Stratum 1-referenced Primary Reference Source in the Central Office.

As for frame relay, the switch will likely provide clock. You need to coordinate with telco to be sure that it has in fact been enabled.