[RP-PPPoE] Problems restarting pppoe

[Graham]

Hi,
     I am having problems with a pppoe link to my ISP that is failing to 
restart after a dropout.

Something appears to lock up at the ISP end that rp-pppoe doesn't clear, 
but a ppp connection from a modem will work fine.

When using my main modem as bridge-only, looking at the packets, I see 
lots of PADI packets, but get no PADO responses from the ISP.
When I plug in a second modem that is configured for PPP, then the 
second modem gets a PPP session very quickly.
If I leave the second modem connected for about 5 minutes, then 
disconnect and reconnect the main modem, then rp-pppoe gets a PADO 
response to its PADI and the connection is re-established (note that the 
main modem does not need to be rebooted, just reconnected to the ADSL line).


Thanks for your assistance.



Some further information...

I am using rp-pppoe 3.11-2 rpm package for Fedora 19.

Getting the first modem to re-sync, or rebooting the first modem does 
nothing to fix the problem, so it doesn't look like a modem problem.

If I only leave the second modem connected for about 30 seconds before 
changing back to the first modem, then the connection still doesn't 
work. I need to leave the second modem connected for about 5 minutes to 
clear the problem.

Stopping rp-pppoe for 5 minutes and then restarting it does not clear 
the problem.

Some searching for issues with my ISP indicates that this fault can 
occur if PAP is used instead of CHAP, however I suspect that this may 
not be an issue until later in the handshake (well after PADI/PADO).