[cisco-nas] bad 5300

[Aaron Leonard]

> Hello All,

> We have several 5300's. recently over a 2 week time period one got 32852
> calls of 1000 bytes or less.  Cu's indicate they cant surf.  This 32852
> is an order of magnitude higher than any other pool we have.

> I have checked the DS0's, the PRI's, all the hardware, sh mo sum, sh mo,
> sh int, everything.  i cant seem to find anything wrong, but at peak this pool
> is getting about 170 connections as opposed to every other pool we have
> at 183.  Im thinking a hex is bad.

> How do I proceed with diagnostics?

> Thanks,

> Marc

OK, so you've done a "show modem" and have found that call hold times
are (statistically speaking) identical across all modems, right?  In 
that case, you do not have a bad DSP (2 adjacent modems) or hex module
(6 adjacent modems.)

Can you catch a "can't surf" call while it is up?  If so, then
please get:

show modem operational s/p
show int async n
show ip route connected
show dialer map

See if you can ping the peer's IP address from the 5300

This sort of "can't surf" behavior is consistent with things like:
- bad or redundant peer routes
- bad dialer map
- multiply allocated client IP address

Aaron