T1 - Input errors, CRCs, and frames increase exponentially over time

[s.wilmes]

I'm having a problem with one out of 12 new T1's. The longer it's up, the more input errors and CRCs it takes. After a shut/no shut it runs pretty clean for 8 to 12 hours. The next day I check it, it's taking many hundred errors per second, even with no load.

I have a 3845 router at the hub and 2621 at the remote. Both have VWIC-2MFT-T1's.

I definitly providing clocking only on one side. SBC (telco) states their connection is clean. I swapped out each router with a spare and problem persists. I use the same config everywhere, so there's something peculiar about this one location.

I've begun to play with the CABLELENGTH cmd on the controller, but it's taking awhile to try all the combinations of parameters. No luck so far...

Any clues??

thanks.

[scottmac]

All sites should be recovering clock from the network (clock-source line or loop).

Unless you are getting cold copper, the network will provide all clocking.

Would the site that you are having problems with also be the one with the clocking set to "Internal?"

You might also check the cabling between your NIU/SmartJack and the router (or CSU/DSU) for damage and proper termination.

Let us know ..

Scott

[s.wilmes]

This problem site and all my other remote locations are CLOCK SOURCE LINE. My hub site has CLOCK SOURCE INTERNAL for all my T1s. My telco (SBC) emphatically claims they do not provide clocking on point-to-point T1s, and it's up to me to do so as I am doing.

I have replaced the patch cables on both sides, even though there was nothing wrong with them. I've replaced every piece of my equipment on both sides, one piece at a time (including my routers), and the problem continues.

If I can trust my telco, then there's something about that location that doesn't like my config that works everywhere else.

thanks

[jroyster]

Definately sounds like a clocking problem.

I'd have the provider loop your CSUs and run intrusive tests. There have been dozens of times a provider tells me "we don't provide clock" and then after numerous tests it is discovered that the circuit was indeed provisioned to provide clock.

For giggles you could set both sides of the T1 to get their clock from the circuit and see if the problem persists.

[joyride_us]

HYi,

I expereicend the same issue. In my case, it was the T1 controller board that was not doing good! I swapped it and everything looked good. In doubt, swap on both sides but basically you should swap only on the side where you get the errors!

Second possility, of course, SBC!

Good luck adn take it easy!

:)