[clug] Troubleshooting PPPOE

[Christopher Giffard]

Hello all! :)

Recently I've been experiencing quite frequent intermittent dropouts  
on my VDSL line - and I'm having trouble diagnosing the cause.

My suspicions currently point to the ISP - the problem initially  
appeared a few weeks ago as a strange inability to view websites which  
were hosted on servers that blocked certain kinds of ICMP traffic,  
like microsoft.com. Setting the MTU on the ppp interface to 1400 fixed  
the issue - only for a short time before it would fail to load those  
sites again, and I'm still not entirely sure why.

This week the connection started to drop out. It wasn't too frequent  
on monday - twice a day - but this morning it's dropped out five or  
six times in the last hour, and I'm manually having to ssh into the  
gateway to restart pppd. I'm sure that there's something wrong with  
the ISP's PPPOE server - the ping latency often peaks at well over a  
second now, where it used to stay fairly solidly at 2 - 10ms. I've  
extracted pages and pages of this kind of thing from syslog:

Apr  7 09:46:14 gateway pppd[16543]: Plugin rp-pppoe.so loaded.
Apr  7 09:46:14 gateway pppd[16543]: pppd 2.4.4 started by root, uid 0
Apr  7 09:46:14 gateway pppd[16543]: PPP session is 6380
Apr  7 09:46:14 gateway pppd[16543]: Using interface ppp0
Apr  7 09:46:14 gateway pppd[16543]: Connect: ppp0 <--> eth1
Apr  7 09:46:14 gateway pppd[16543]: PAP authentication succeeded
Apr  7 09:46:14 gateway pppd[16543]: peer from calling number  
00:30:88:10:E6:91 authorized
Apr  7 09:46:14 gateway pppd[16543]: not replacing existing default  
route via 192.168.0.1
Apr  7 09:46:14 gateway pppd[16543]: Cannot determine ethernet address  
for proxy ARP
Apr  7 09:46:14 gateway pppd[16543]: local  IP address 123.123.123.123
Apr  7 09:46:14 gateway pppd[16543]: remote IP address 123.123.123.123
Apr  7 10:21:04 gateway pppd[16543]: Hangup (SIGHUP)
Apr  7 10:21:04 gateway pppd[16543]: Connect time 34.9 minutes.
Apr  7 10:21:04 gateway pppd[16543]: Sent 130378265 bytes, received  
31249965 bytes.
Apr  7 10:21:04 gateway pppd[16543]: Connection terminated.

Because of the constant hangups, I thought it might have been the  
modem, but we replaced it, and we're still experiencing the dropouts.  
Often, though, pppd doesn't terminate the connection - the connection  
just stops working, and I have to manually restart it.

Essentially, can anyone think of anything at all on the local machine  
which could possibly be causing to repeatedly drop the connection? I'd  
like to be absolutely sure that it isn't a problem with our gateway  
server before I whine to my ISP.

The modem is a (now new) TransACT Marconi, we're on a 10Mbit down/  
1Mbit up VDSL line - connected to an ASUS Gigabit ethernet card. The  
technician checked the line and gave it the all clear, too - so it  
isn't that.

I would be incredibly grateful for any assistance - I'm stumped!

Regards,
Christopher Giffard