PPPoE won't reconnect after link loss

clarknova

Attached are four files. Here's how I got them:

1. Boot system fresh. PPPoE connected.
2.```

clog /var/log/system.log | egrep '(mpd|ppp)' > ppp_working.txt
cp /var/etc/mpd_wan.conf mpd_wan_working.conf.txt

3\. Pull plug on WAN NIC. Wait 2 minutes. Re-plug. (WAN not connecting)
4.```

clog /var/log/system.log | egrep '(mpd|ppp)' > ppp_not_working.txt
cp /var/etc/mpd_wan.conf mpd_wan_not_working.conf.txt

5. Reboot to working WAN.

Null Service Name is checked.
mpd_wan_not_working.conf.txt
mpd_wan_working.conf.txt
ppp_not_working.txt
ppp_working.txt

clarknova

Reproduceable in 2.0-RELEASE amd64. Haven't tested other arches.

Opened a bug here: http://redmine.pfsense.org/issues/1943

jimp

Try this:

ngctl list > /root/ngctl-before.txt

unplug, replug, wait for it to try reconnecting

Then run:

ngctl list > /root/ngctl-after.txt
diff -ub /root/ngctl-before.txt /root/ngctl-after.txt

I'm curious if there are any differences when it works and when it doesn't.

clarknova

em0_vlan20X = mlppp member interface
em1_vlanY (where Y is 7 or 666) = disabled static interface

ngctl-before.txt
ngctl-after.txt
diff.txt

jimp

Does your pppoe0 interface still exist when it's trying to reconnect?

For curiosity sake, while it's not working, go to Diag > Command, in the PHP box, and try:

pfSense_ngctl_attach(".", "ng0");

Or perhaps

pfSense_ngctl_attach("ng0", "pppoe0");

Failing that, give this a try: Edit usr/local/sbin/ppp-linkdown and comment out or remove this line (keep an unaltered copy around though so you can restore the copy if it doesn't help)

/usr/sbin/ngctl shutdown $1:

clarknova

I'll post my updates here: http://forum.pfsense.org/index.php/topic,41061.msg225361.html#msg225361, as that seems to be the current thread for this topic.

Craigusoz

Hi, I'm looking to set up pfSense in a small business environment, but am very concerned about this apparent problem that seems to have no fix. Does this occur universally, or just with some ISP's ?

wallabybob

I believe some changes went into pfSense 2.0.1 to address this problem. Its not universal. I am not alone in being able to report that my pppoe successfully restarts after link loss. I believe at least one of the pfSense developers reported they were unable to reproduce the problem.

jimp

Several of us (developers) use PPPoE at home and various other places, and have never been able to reproduce this as far as I'm aware.

It would seem to hit certain ISPs more than others.

Craigusoz

OK, that is somewhat reassuring, thanks for taking the time to reply.

minimaster

I've Pfsense 2.0 and lately 2.01 running on Alix 2D13. vr1 with a static VDSL connection from a seperate VDSL router and vr2 with a PPPoE connection via Teledat430/T-DSL connection. Same issue here with both 2.0 and 2.01; when the DSL cable is disconnected >2 minutes it won't reconnect. Must reboot the Alix to get the PPPoE connection alive again.
Snippet from the log:
Jan 9 20:10:57 ppp: [opt1_link0] PPPoE connection timeout after 9 seconds
Jan 9 20:10:48 ppp: [opt1_link0] PPPoE: Connecting to ''
Jan 9 20:10:48 ppp: [opt1_link0] Link: reconnection attempt 87
Jan 9 20:10:44 ppp: [opt1_link0] Link: reconnection attempt 87 in 4 seconds
Jan 9 20:10:44 ppp: [opt1_link0] LCP: Down event
Jan 9 20:10:44 ppp: [opt1_link0] Link: DOWN event
Jan 9 20:10:44 ppp: [opt1_link0] PPPoE connection timeout after 9 seconds
Jan 9 20:10:35 ppp: [opt1_link0] PPPoE: Connecting to ''
Jan 9 20:10:35 ppp: [opt1_link0] Link: reconnection attempt 86
Jan 9 20:10:34 ppp: [opt1_link0] Link: reconnection attempt 86 in 1 seconds

clarknova

Have you tried this: http://forum.pfsense.org/index.php/topic,40671.msg219480.html#msg219480

It worked for me in 2.0. I haven't tested since upgrading to 2.0.1 though.

minimaster

Tried all three suggestions, including reboot after the file edit. Still if the PPPoE line is being disconnected for a longer while it won't reconnect. The kernel recognizes on the interface the reconnected modem:
Jan 9 21:44:46 kernel: vr2: link state changed to UP
Jan 9 21:44:46 check_reload_status: Linkup starting vr2

, but ppp continues with endless attempts and timeouts
Jan 9 21:45:05 ppp: [opt1_link0] LCP: Down event
Jan 9 21:45:05 ppp: [opt1_link0] Link: DOWN event
Jan 9 21:45:05 ppp: [opt1_link0] PPPoE connection timeout after 9 seconds
Jan 9 21:44:56 ppp: [opt1_link0] PPPoE: Connecting to ''
Jan 9 21:44:56 ppp: [opt1_link0] Link: reconnection attempt 33
Jan 9 21:44:52 ppp: [opt1_link0] Link: reconnection attempt 33 in 4 seconds

wallabybob

It could be worthwhile doing a packet capture on the physical interface. Does the capture show outgoing PPP startup attempts? Does the capture show incoming responses? If not, might be time to call your ISP's tech support.

minimaster

A packet capture does not show any packets (not a single one) on the interface while the log shows all the attempts and timeouts.

wallabybob

@minimaster:

A packet capture does not show any packets (not a single one) on the interface while the log shows all the attempts and timeouts.

Interesting! Packet capture on vr2? or pppoe0?

minimaster

Its a capture on PPPoE1. PPPoE0 is is the name for an old Vlan. Actually I deleted PPPoE0 a while ago. PPPoE1 did remain and is assigned to vr2.
Actually I've now deleted the interface (vr2) and the PPP and recreated/assigned vr2 to new PPPoE type interface (now named PPPoE0), but no change; after a longer disconnect the PPPoE interface won't reconnect.

wallabybob

I wonder if ppp is stuck waiting (possibly erroneously) for a

Jan 10 10:02:07 pfSense ppp: [wan_link0] Link: UP event

. This is part of a bigger sequence of events logged when I disconnected my ppp link and then reconnected it:

Jan 10 10:02:10 pfSense ppp: web: web is not running
Jan 10 10:02:10 pfSense ppp: [wan] Bundle: Interface ng0 created
Jan 10 10:02:10 pfSense ppp: [wan_link0] Link: OPEN event
Jan 10 10:02:10 pfSense ppp: [wan_link0] LCP: Open event
Jan 10 10:02:10 pfSense ppp: [wan_link0] LCP: state change Initial –> Starting
Jan 10 10:02:10 pfSense ppp: [wan_link0] LCP: LayerStart
Jan 10 10:02:10 pfSense ppp: [wan_link0] PPPoE: Connecting to ''
Jan 10 10:02:10 pfSense ppp: PPPoE: rec'd ACNAME "pe5.dc1-vars"
Jan 10 10:02:10 pfSense ppp: [wan_link0] rec'd proto LCP while dead
Jan 10 10:02:10 pfSense ppp: [wan_link0] PPPoE: connection successful
Jan 10 10:02:10 pfSense ppp: [wan_link0] Link: UP event
Jan 10 10:02:10 pfSense ppp: [wan_link0] LCP: Up event
Jan 10 10:02:10 pfSense ppp: [wan_link0] LCP: state change Starting –> Req-Sent
Jan 10 10:02:10 pfSense ppp: [wan_link0] LCP: SendConfigReq #1
Jan 10 10:02:10 pfSense ppp: [wan_link0] PROTOCOMP
Jan 10 10:02:10 pfSense ppp: [wan_link0] MRU 1492
Jan 10 10:02:10 pfSense ppp: [wan_link0] MAGICNUM 82cb834a
Jan 10 10:02:10 pfSense ppp: [wan_link0] LCP: rec'd Configure Ack #1 (Req-Sent)
Jan 10 10:02:10 pfSense ppp: [wan_link0] PROTOCOMP
Jan 10 10:02:10 pfSense ppp: [wan_link0] MRU 1492
Jan 10 10:02:10 pfSense ppp: [wan_link0] MAGICNUM 82cb834a
Jan 10 10:02:10 pfSense ppp: [wan_link0] LCP: state change Req-Sent –> Ack-Rcvd
Jan 10 10:02:12 pfSense ppp: [wan_link0] LCP: rec'd Configure Request #2 (Ack-Rcvd)

I wonder if the empty packet capture indicates ppp was waiting for the other end to "speak" and the other end was waiting for pfSense ppp to speak. The extract from my log suggests to me that my ISP spoke before my pfsense ppp did.

xbipin

what was the solution to this coz same issue for me since a very long time now