I have a situation where an entire MLPPP bundle goes down because of a single T1 taking errors. My thought is that when the errors are severe enough, the result is lost mlppp 'control' packets that cause the whole bundle to crash. Obviously, when this occurs, the normal course is to shut the offending T-1, troubleshoot with the LEC and bring it back up only when it's fixed. However, I'm not always available and I would prefer if there was a way for the router to sense that a T-1 is taking too many CRC, input errors, (via a threshold?) and automatically put the interface into a down/down state. Then, the rest of the bundle would survive until I had time to fix.
Has anyone tried this before? Is there any ability to do it?
Any ideas appreciated.