PPPoE authentication issue with ISP

Installing and Using OpenWrt
1

Hello friends,
Need your comments on this issue.
I am using the version: DESIGNATED DRIVER (Bleeding Edge, 12009)
I am using PPPoE setting with the ISP ( ACT fibernet, a large player in India).
It was working all good for past 2-3 years.
Suddenly the PPPoE authentication has started failing.
ISP tech support showed with a TP-Link stock FW router that it works; so I have nothing to blame them.
Also checked in DHCP mode, it works. So no issue with the physical port.

Check the log below; same thing keeps repeating.

Sep 14 13:49:36  pppd[5009]: pppd 2.4.7 started by root, uid 0
Sep 14 13:49:36  pppd[5009]: PPP session is 56541
Sep 14 13:49:36  pppd[5009]: Connected to 44:46:fe:46:66:c1 via interface eth0
Sep 14 13:49:36  kernel: [10364.865343] pppoe-wan: renamed from ppp0
Sep 14 13:49:36  pppd[5009]: Using interface pppoe-wan
Sep 14 13:49:36  pppd[5009]: Connect: pppoe-wan <--> eth0
Sep 14 13:49:39  pppd[5009]: LCP terminated by peer
Sep 14 13:49:39  pppd[5009]: Modem hangup
Sep 14 13:49:39  pppd[5009]: Connection terminated.
Sep 14 13:49:39  pppd[5009]: Sent PADT
Sep 14 13:49:39  pppd[5009]: Exit.
Sep 14 13:49:39  netifd: Interface 'wan' is now down
Sep 14 13:49:39  kernel: [10368.096904] eth0: link down
Sep 14 13:49:39  netifd: Interface 'wan' is disabled
Sep 14 13:49:39  netifd: Interface 'wan' is enabled
Sep 14 13:49:39  netifd: Interface 'wan' is setting up now
Sep 14 13:49:39  kernel: [10368.108229] IPv6: ADDRCONF(NETDEV_UP): eth0: link is not ready
Sep 14 13:49:40  netifd: Network device 'eth0' link is down
Sep 14 13:49:40  netifd: Interface 'wan' has link connectivity loss
Sep 14 13:49:40  netifd: wan (5055): Command failed: Permission denied
Sep 14 13:49:41  netifd: Network device 'eth0' link is up
Sep 14 13:49:41  netifd: Interface 'wan' has link connectivity
Sep 14 13:49:41  netifd: Interface 'wan' is setting up now
Sep 14 13:49:41  kernel: [10369.721954] eth0: link up (1000Mbps/Full duplex)
Sep 14 13:49:41  pppd[5084]: Plugin rp-pppoe.so loaded.
Sep 14 13:49:41  pppd[5084]: RP-PPPoE plugin version 3.8p compiled against pppd 2.4.7
Sep 14 13:49:41  pppd[5084]: pppd 2.4.7 started by root, uid 0
Sep 14 13:49:41  pppd[5084]: PPP session is 32408
Sep 14 13:49:41  pppd[5084]: Connected to 84:46:fe:26:c6:c0 via interface eth0
Sep 14 13:49:41  kernel: [10369.876109] pppoe-wan: renamed from ppp0
Sep 14 13:49:41  pppd[5084]: Using interface pppoe-wan
Sep 14 13:49:41  pppd[5084]: Connect: pppoe-wan <--> eth0
Sep 14 13:49:44  pppd[5084]: LCP terminated by peer
Sep 14 13:49:44  pppd[5084]: Modem hangup
Sep 14 13:49:44  pppd[5084]: Connection terminated.
Sep 14 13:49:44  pppd[5084]: Sent PADT
Sep 14 13:49:44  pppd[5084]: Exit.
2

Rskesh
That's an ancient version i think......have you flashed more recent ones?

1 Like
3

It's 17.x ; old but not that old.

Also the pppd version is the same even on 18.x; there is no update.

I am yet to test the problem on 18.x box but it seems unlikely to have any impact.

 OpenWrt 18.06.0
 ---------------------------
root# pppd --version
pppd version 2.4.7


 DESIGNATED DRIVER (Bleeding Edge, 12009)
 -----------------------------------------------------
root# pppd --version
pppd version 2.4.7
4

Also posting related configs. I had tweaked some of them from default to make it work smoothly. This config did work for the past 3 years until now. Something has changed on the ISP side of the protocol. I don't understand PPPoE enough to even start debugging it.

root# cat /etc/ppp/options
#debug
logfile /dev/null
noipdefault
noaccomp
nopcomp
nocrtscts
lock
maxfail 0
lcp-echo-failure 0
lcp-echo-interval 1
5

Try disabling IPv6, like this in /etc/config/network:

config interface 'wan'
	option proto 'pppoe'
	option ipv6 '0'
	option pppd_options 'noipv6'
	option username 'xxxxx'
	option password 'xxxxx'
        [...]
1 Like
6

I already do : option ipv6 '0'
Do you think pppd_options 'noipv6' will have an impact?

8

Please uncomment debug and post detailed logs. It should tell us which end is initiating disconnect.

1 Like
9

That's a good pointer. Will definitely give it a try and post the outcome here.

10

Huawei modem?

Have you got control of this modem

11

The first line by itself is not sufficient to fully disable IPv6, ironically. You'll find that if you search the forum for PPPoE & IPv6. My reconnection issues only got fixed with the pppd_options setting.

Wouldn't hurt to upgrade your old setup to a new version though, 21.02 just went stable and what you're using is Swiss cheese now, security wise.

1 Like
12

There is no modem.
The ISP just provides a CAT6 cable end-point which I plug into my x86 device running OpenWRT.

13

So I tried 19.07.8 and it worked. There are 2 things different though:

  1. It's a different HW (but same wan interface name eth0)
  2. The /etc/ppp/options options are default.

So I have got a working baseline; will proceed from here.

Thanks to all of you who commented on this and tried to help in your capacity.

1 Like
14

I believe the abrupt "terminated by peer" means that your client and the ISP could not negotiate a mutually compatible protocol for compression / encryption / etc. Better logging is needed to see exactly what happened.

15

Funny I was creating a new topic for the exact same issue, I notice it's cropping up again.
As soon as i reboot the router, the link comes up. Works for a few hours, sometimes even days. And the story repeats.
It's so predictable that I have created a cron, if google dns is not reachable over this ISP, reboot the router. Runs every 1 hour.

I understand I am running an old OpenWRT OS. But i have seen no update in the pppoe package.

At this point I will be happy to find a better work-around.
Reboot is 2 minute impact.
Can i reduce it to 30s may be?

16

What device are you using?

17

TP-Link 1043ND.

18

I see one other related thread, where "ip route" was getting removed.
In my case, I find that the output of "ip route show" is identical to the working state.