On 2020-01-03, jrmu <j...@ircnow.org> wrote: > Greetings, > > I am getting LCP keepalive timeout errors when setting up a PPPOE > connection to a fiber optic line. I am attempting this on > BeagleBone Black running OpenBSD 6.6 GENERIC armv7. > > I have two interfaces which I set up as follows: > > # cat /etc/hostname.cpsw0 > up > > # cat /etc/hostname.pppoe0 > inet 0.0.0.0 255.255.255.255 NONE \ > pppoedev cpsw0 authproto pap \ > authname '12345...@isp.net' authkey 'abcd1234' up > dest 0.0.0.1 > #inet6 eui64 > !/sbin/route add default -ifp pppoe0 0.0.0.1 > #!/sbin/route add -inet6 default -ifp pppoe0 fe80::%pppoe0 > > (IPv6 is not supported by my ISP) > > Internet works on bootup for a while, but after about 5 minutes, > I see this error in /var/log/messages: > > /bsd: pppoe0: LCP keepalive timeout > > Networking will work for about 5 minutes, then there will be a > disconnection lasting for about 30 seconds. Each time, it's because > of an LCP keepalive timeout.
Have a look at what's being sent on cpsw0 (tcpdump etc), is the ISP responding to your LCPs at all? Possible workaround: try keeping some "user" traffic running (a ping would probably do it). pppoe(4)/sppp only uses LCP keepalives if there isn't other traffic. > I can confirm that similar PPPOE setups as above used to work on > my x64 PCs until last month, when I suspect my ISP changed its > settings. My non-OpenBSD router continues to function as normal using > the fiber optic PPPOE. I am convinced that I need to modify the > LCP keepalive settings to match my ISP's. It's not a case of "matching the ISP", if they receive an LCP echo request they are supposed to respond to it (and likewise if they send you an LCP echo request you are supposed to respond), there is no matching of timers involved. > Someone had suggested I take a look into npppd. npppd won't help, it is "server side" only. I would: 1. confirm if the problem also exists on amd64 or some other more arch, to make sure it's not an OpenBSD/arm-specific problem 2. if possible (tcpdump on the device itself, switch span port, ethernet hub, separate box with 2x ethernet ports bridged together where you can run tcpdump, etc) see what is different about the packets transmitted on the ethernet interface on OpenBSD compared to your non-OpenBSD router. 3. maybe ask your ISP what they changed? (whether you will get a useful response to this depends on your ISP of course... I try to avoid ISPs where there's no way to get in contact with people who can answer such questions if possible though!)
