VWIC-2MFT-T1 in WIC slot, PPP Multilink and Slip Sec

[Erick Bergquist]

Hi,

Have a strange issue where we have 2 3745's at head end with 3 VWIC-2MFT-T1 in chassis WIC slots. 6 T1s total at head end. 4 Remote sites - 2620XM or 2650XM with a single VWIC-2MFT-T1 in the chassis WIC slot.

Each 3745 has a connection to 2 sites. All the connections are 2 T1s PPP Multilinked together.

On the remote routers, on only the 2nd T1 we are seeing Slip and errored seconds. Both 3745's are set for internal clocking for all T1s and remote routers are set to line clocking.

We have no AIM modules in any of the routers.

Any idea why we are getting slip and errored seconds? We have had these T1s checked out multiple times. Data and voice is passing fine across these, but would like errors to go away.

Also, we are seeing input errors on the logical serial interfaces only during day when data is passing over T1s. Afterhours the input errors go away which leads me to believe that these are related to data/traffic. They increase slowly and not rapidly.

The T1 controller is only reporting slips on 2nd T1 at remotoes. Thats it for controller errors. Both serials see the input errors slowly.

Sample:

Data in current interval (43 seconds elapsed):

0 Line Code Violations, 0 Path Code Violations

4 Slip Secs, 0 Fr Loss Secs, 0 Line Err Secs, 0 Degraded Mins

4 Errored Secs, 0 Bursty Err Secs, 0 Severely Err Secs, 0 Unavail Secs

Total Data (last 17 15 minute intervals):

0 Line Code Violations, 0 Path Code Violations,

1285 Slip Secs, 0 Fr Loss Secs, 0 Line Err Secs, 0 Degraded Mins,

1285 Errored Secs, 0 Bursty Err Secs, 0 Severely Err Secs, 0 Unavail Secs

68322 input errors, 12474 CRC, 51177 frame, 0 overrun, 0 ignored, 4671 abort

[Erick Bergquist]

I've been doing some looking into this, and found a usenet posting that indicates the VWIC-2MFT-T1 has a single clock chip which could be a issue in the config this is in perhaps.

I'm not able to find anything yet on Cisco or CCO stating this.

The 2 T1s at remote site go into same VWIC-2MFT-T1 but at the head end site 1 T1 goes into one VWIC-2MFT-T1 and the other T1 goes into another VWIC-2MFT-T1. Head end is set to all internal clocking.

So thus the remote site set fo line clocking, is getting different clocks on each T1 perhaps if the VWIC-2MFT-T1 only has a single clock. Does it prefer the 0 port for line clocking? Thats why the 2nd port see's slips?

Anyone else run into this?

[averzosa]

Hi Erick,

Good Day! :-)

When both ports on a VWIC-2MFT are configed as clock source line, then clock is being provided by the remote ends and it must ultimately be sourced by the same timing reference. Otherwise slips will occur. You choices are to either 1)reconfigure one T1 line such the router provides clock (clock source internal), or 2) move a T1 to a different card.

But note- if you are using NM-xFE2W cards, the VWIC-MFT ports configured as clock source line require that all ports be timed by the same reference. This restriction does not apply to the WIC-T1-DSU.

I hope this information helps. :-)

Best Regards,

Ansley R. Verzosa

[Erick Bergquist]

I found a cisco document that states that with the VWIC-2MFT that there is one on-board PLL chip (Phase Loop Lock). Doc saids that the card can receive clock on one port and use that clock to drive a clock on the other port.

We are only seeing slips on remote sites with a single VWIC-2MFT in a 2600 series WIC slot on 2nd port. Both ports are in use with connection back to head end.

Head end is 3745 with 2 VWIC-2MFT in WIC slots and a T1 comes into each card. head end has internal clocking configured on all ports. We have 2 clock sources since it is 2 VWICs.

I think the reason remote end is seeing slip secs is because both the T1s are coming from different VWICs at head end and the VWIC has a single PLL chip. Port 0 has clock from line because it is first on card and port 1 doesn't have good clock so thats why wee see slips just on 2nd T1 at remote.

I have suggested to move both T1s to same VWIC at both sides but still wonder if slip secs will coninue due to the single PLL chip. Or if we should configure one port as internal and one as line at both ends to get error free perhaps.

If I configure internal clocking on a remote site VWIC on port 1 then we see slip secs on port 1 on the VWIC on 3745 side which is going to another remote site.

So it seems if using VWIC-2MFTs one has to really plan out their clock sources, to avoid a possible cascading problem where you could effect another T1.

Does anyone know if the VWIC-2MFTs are meant to do PPP Multilink, or 2 T1s for data transport? I think they are mainly a voice card that can also do data but data wasn't main focus.

[thisisshanky]

I have a similar setup for one my customers...with 3745 at one end and a 2600XM at other end.

3745 has two NM-HDV's with one VWIC-2MFT-T1s on each NM-HDV. Each VWIC terminates a data T1 and a voice T1. The two data T1s run back to the 2600 XM which has a NM-HDV with a VWIC-1MFT-T1 (1 port T1) and a WIC-1DSU-T1.

I know this configuration is weird but it was done to save money for the customer and use whatever hardware we had.

I have done a multilink on these two T1s. I dont see any clock slips or errors on the both the t1s. But the biggest problem we are having is that the 3745 crashes as soon as we transfer a file across the 3 Mb pipe.

We tried upgrading/downgrading IOS versions..to the latest and greatest Cisco IOS versions..but with no results..

We tried swapping the 3745 with a brand new 3745 with NM-HDVs and T1 wics...no results on this either..

Have a Cisco TAC case open on this one...Cisco development team is researching on the crash files..I dont know what to do next...

Could this be because I have a 2 VWICs at one end and 1 VWIC and 1WIC-DSU-T1 at the other end?? Any compatibility issues ??

I am yet to try installing two VWICs at either ends.I will try this some time next week..

Meanwhile..let me know what you guys think about this problem..

3745 NM-HDV1--VWIC========T1======VWIC---NM-HDV1-2600

|__NM-HDV2--VWIC========T1======WIC-1DSU-T1-----|

[Erick Bergquist]

I don't see anything wrong with that at the moment. Have you tried to do just equal cost load balancing with the T1s not in a bundle to isolate it down to a PPP mlink issue?

Keep us posted.

[thisisshanky]

I have tried that. Without the bundle, there are no issues. It is definitely something to do with PPP multilink. I also think its related to the combination of the hardware that I am using. If this is a software bug, its some thing that has not yet been fixed in Cisco IOS...