12016/GRP sending LCP reconfigs on established Multilink PPP session

All:

I'm trying to install a pair of T1s using Multilink PPP to connect my
Linux + Sangoma Wanpipe firewall/router to my ISP's Cisco 12016/GRP
running v12.0 of (I assume) IOS.

The PPP sessions all establish just fine, but anywhere from thirty seconds
to ten minutes from the initial connect, and occurring fairly frequently
thereafter, the Cisco starts sending LCP ConfReqs with a new magic number.
This tears down the existing PPP link and rebuilds it.

This is causing me problems due to software issues on my end that I am
working on. However, even *without* my software problem killing the link,
having these constant reconfigures is annoying - how much of my bandwidth
is going to be lost to reconfiguring the PPP link every few minutes?

Can anybody tell me WHY the Cisco wants to reconfigure the link so much?
And is there any way to turn that behavior off? My ISP's escalated support
rep spent about an hour poking around in configuration options trying to
find something to control it and couldn't.

Thanks in advance for any help.

 

I have fixed this. It appears to be a bug in Cisco's PPP code. Please see
my other post in this group.

--
John Hardin
Internal Systems Administrator (Seattle)
CRS Retail Systems, Inc.
3400 188th Street SW, Suite 185
Lynnwood, WA 98037
voice: (425) 672-1304
fax: (425) 672-0192
email:
web: http://www.crsretail.com