I work for a mid-sized company on Norwalk, CT. Our main office is linked to our 2nd largest office 50 miles away via two bonded P2P T-1 circuits. Both ends are linked via Cisco 2621 routers.
One of the bonded T-1's has, over the past 3 months, started going up and down multiple times (anywhere from 5 to 20 times) every 14-20 days. The time down is anywhere from 45 seconds to 20 minutes. This particular circuit has been in place for close to 7 years and we NEVER had any issues with it.
Now - after going around and around with the provider (AT&T), they are telling us that since it's a P2P circuit, they don't provide any timing on the circuit and we should have the interface at one end set to "internal" and the other end set to "line". I have a Cisco consultant who is very knowledgable and experienced and he thinks they are full of it. Basically - here is what I would like to know:
1) Is it Cisco SOP (Standard Operating Procedure) or Best Practice to enable internal clocking on one end of a P2P circuit?
2) Why would this circuit never have any problems and then all of a sudden we need to set internal clocking? My wory is that AT&T changed a bunch of stuff on their end which caused this and we're having to adapt to compensate - that's fine - until they rectify the problem and we end up with more issues.
3) The biggest issue I have is that if it's best practice or SOP - shouldn;t we enable internal clocking on one side of ALL our P2P circuits? If it's one way or another, great - I just don't like having to do it on one but not on another.
4) Lastly - if this IS something Cisco reccommends, where could I get a copy of that best practice documentation?
Hopefully this is something someone can answer...I'm sure a lot of you hate dealing with telecom providers as much as I do.
Perhaps I misinterpreted your question about all circuits. Especially for your other P2P circuits I recognize that you have a difficult choice. I generally adhere to the adage that if it is not broken, then do not fix it/change it. On that basis I would be inclined to let the circuits that are currently working ok go on as they are, but being prepared to configure clocking internal if any of them start to develop instability.
On the other hand I can also see the point that if the provider is not providing clocking then it is likely that at some point they may become unstable and that it might be good to be proactive configure clocking.