PPPoE reconenction fix - mpd fix ($100)

eri--

Can you do me a 'ngctl list' when in this situation?

xbipin

here is the output before and after

when normally connected after reboot


$ ngctl list
There are 17 total nodes:
  Name: <unnamed>Type: socket          ID: 0000001c   Num hooks: 0
  Name: <unnamed>Type: socket          ID: 0000001b   Num hooks: 0
  Name: <unnamed>Type: pppoe           ID: 00000011   Num hooks: 2
  Name: <unnamed>Type: socket          ID: 00000010   Num hooks: 0
  Name: ng0             Type: iface           ID: 0000000d   Num hooks: 1
  Name: mpd10888-wan_link0-lt Type: tee             ID: 0000000f   Num hooks: 2
  Name: vr1             Type: ether           ID: 00000002   Num hooks: 1
  Name: vr2             Type: ether           ID: 00000003   Num hooks: 0
  Name: bridge0         Type: ether           ID: 00000018   Num hooks: 0
  Name: mpd10888-stats  Type: socket          ID: 00000012   Num hooks: 0
  Name: mpd10888-cso    Type: socket          ID: 0000000b   Num hooks: 0
  Name: mpd10888-wan-mss Type: tcpmss          ID: 00000013   Num hooks: 2
  Name: mpd10888-wan    Type: ppp             ID: 0000000e   Num hooks: 3
  Name: mpd10888-eso    Type: socket          ID: 0000000c   Num hooks: 0
  Name: mpd10888-lso    Type: socket          ID: 0000000a   Num hooks: 1
  Name: wlan0           Type: ether           ID: 00000017   Num hooks: 0
  Name: ngctl4052       Type: socket          ID: 00000023   Num hooks: 0</unnamed></unnamed></unnamed></unnamed>

when endless loop of reconenction


$ ngctl list
There are 16 total nodes:
  Name: <unnamed>Type: socket          ID: 0000001c   Num hooks: 0
  Name: <unnamed>Type: socket          ID: 0000001b   Num hooks: 0
  Name: <unnamed>Type: pppoe           ID: 00000011   Num hooks: 2
  Name: <unnamed>Type: socket          ID: 00000010   Num hooks: 0
  Name: ng0             Type: iface           ID: 0000000d   Num hooks: 0
  Name: mpd10888-wan_link0-lt Type: tee             ID: 00000028   Num hooks: 2
  Name: vr1             Type: ether           ID: 00000002   Num hooks: 1
  Name: vr2             Type: ether           ID: 00000003   Num hooks: 0
  Name: ngctl23563      Type: socket          ID: 0000002d   Num hooks: 0
  Name: bridge0         Type: ether           ID: 00000018   Num hooks: 0
  Name: mpd10888-stats  Type: socket          ID: 00000012   Num hooks: 0
  Name: mpd10888-cso    Type: socket          ID: 0000000b   Num hooks: 0
  Name: mpd10888-wan    Type: ppp             ID: 0000000e   Num hooks: 1
  Name: mpd10888-eso    Type: socket          ID: 0000000c   Num hooks: 0
  Name: mpd10888-lso    Type: socket          ID: 0000000a   Num hooks: 2
  Name: wlan0           Type: ether           ID: 00000017   Num hooks: 0</unnamed></unnamed></unnamed></unnamed>

xbipin

@ermal:

Can you do me a 'ngctl list' when in this situation?

any luck?

xbipin

still no fix for this

dhatz

Have your tried pfsense 2.1-DEVEL just in case the latest mpd 5.6 improves things?

xbipin

havent tried it coz its not much in a usable state currently for light production, mayb later

dhatz

Try adding

set bundle no noretry

src: http://unix.derkeiler.com/Mailing-Lists/FreeBSD/questions/2007-09/msg01895.html

Apparently it was also discussed here a couple of years ago:

According to http://mpd.sourceforge.net/doc5/mpd4.html

Changes since version 4.1rc2:
Bundle's noretry option is enabled by default now.

xbipin

i tried to feed that line in using edit file from the web GUI but it seems pfsense seems to remove that entry if rebooted or even if i goto interfaces and click disconnect and then connect

any help regarding some php code change to keep that entry intact?

xbipin

i managed to edit the interfaces.inc so it adds the line which it does now but i guess mpd ignores it as the system log gives the below message

ppp: mpd_wan.conf:31: Incorrect context for: 'set bundle no noretry'

xbipin

actually the command to use was this

set link no noretry

it didnt help coz the original issue which u mentioned in that first link was his connection wasnt redialing that was due to noretry defaulting to yes but that was in mpd4 and i guess pfsense uses mpd5 in v2.0.1 so thats y that command isnt recognized, only max-redial is the command available which cant be set to some number or 0 for unlimited and -1 to disable redial which pfsense defaults it to 0 so my connection is actually redialed but due to some other reason and the isp's fiber optic modem or their bras server, mpd isnt able to reconnect untill its rebooted so obviously something different happens during pfsense reboot time connection or mayb because the vr1 resets which then allows pfsense to connect fine on reboot but to replicate that i removed the network cable as well and tried and try replugging it in but doesnt help.

dhatz

xbipin,

Let me try to recap:

You had no problem with PPPoE auto-reconnection with pfsense 1.2.3
You only started having problems with 2.0-something (BETA, RC etc)
You are able to reproduce the problem at will
There are others who have a similar problem, but most pfsense users don't

I think your best option would be to obtain logs with full verbosity and post them along your mpd config files to the mpd-forum at http://sourceforge.net/projects/mpd/forums/forum/44692

xbipin

i already did that in their mailing list, hardly got any reply at all and there r many others with the problem but they r either using 1.2.3 or trying to live with 2.0.1 like me. the basic reconnection issue is mpd attempts but is unable to reconnect and if u read the past posts, i have provided logs, config etc etc almost more than 50 times now till date.

bytheway, all routers available in the market etc work very well over the same link, then y does mpd have issues?

dhatz

@xbipin:

there r many others with the problem but they r either using 1.2.3 or trying to live with 2.0.1 like me.

Then why aren't any others bidding here to try to get it fixed?

xbipin

coz officially ermal was on this to get it solved but then mayb he realized its not a easy fix by simply patching some php code so this thread went old as i was out of town and got abandoned so most might not have read it also now but if u look at other threads here and there in the complete forum then there r quiet a few ppl suffering the same and it was already discussed many many times but didnt yield any permanent solution.

after a lot of tracing in the past also it came to light that there was something to do with the isp BRAS or pppoe server or whatever they call it and some security measures but my point is that how do commercial routers overcome this and pfsense or mpd doesnt

w0w

Looks like I've run into the same issue, not absolutely the same, but similar and solution for the problem is the same - reboot of pfsense always helps.
Is there the way to reboot pfsense automatically after XX PPPoE connection retries? I know this is not the best way, but at least it works always for me.

eri--

not really i backed off just other priorities first.

Sorry about that.

w0w

ermal, at least you are monitoring ::)

Looks like I've found the solution, that works at least for me
I've added two lines to the end of ppp-linkdown

ifconfig em0 down
ifconfig em0 up

And looks like reconnection bug is gone.
P.S. em0 is interface, where PPPoE is
P.P.S. I think it's related to driver.

xbipin

@w0w:

ermal, at least you are monitoring ::)

Looks like I've found the solution, that works at least for me
I've added two lines to the end of ppp-linkdown

ifconfig em0 down
ifconfig em0 up

And looks like reconnection bug is gone.
P.S. em0 is interface, where PPPoE is
P.P.S. I think it's related to driver.

can u mention the location of the file and where exactly did u put those lines, ill give it a try too

eri--

Certainly that will reset the device by doing up/down.
Just not a generic solution :)

w0w

@xbipin:

@w0w:

ermal, at least you are monitoring ::)

Looks like I've found the solution, that works at least for me
I've added two lines to the end of ppp-linkdown

ifconfig em0 down
ifconfig em0 up

And looks like reconnection bug is gone.
P.S. em0 is interface, where PPPoE is
P.P.S. I think it's related to driver.

can u mention the location of the file and where exactly did u put those lines, ill give it a try too

usr/local/sbin

You can also use Commands-Find File in WinSCP if you have windows


#!/bin/sh
if [ -f /tmp/$1up ] && [ -f /conf/$1.log ]; then
        seconds=$((`date -j +%s` - `/usr/bin/stat -f %m /tmp/$1up`))
        /usr/local/sbin/ppp-log-uptime.sh $seconds $1 &
fi
if [ "$3" != "" ]; then
	echo "Removing states from $3" | logger -t ppp-linkdown
	/sbin/pfctl -k 0.0.0.0/0 -k $3/32
	/sbin/pfctl -k $3/32
	pfctl -K $3/32
fi
if [ "$4" != "" ]; then
	echo "Removing states to $4" | logger -t ppp-linkdown
	/sbin/pfctl -b 0.0.0.0/32 -b $4/32

	if [ -f "/tmp/${interface}_defaultgw" ]; then
		route delete default $4
	fi

fi
# delete the node just in case mpd cannot do that
/usr/sbin/ngctl shutdown $1:
if [ -f "/var/etc/nameserver_$1" ]; then
	# Remove old entries
	for nameserver in `cat /var/etc/nameserver_$1`; do
		/sbin/route delete $nameserver >/dev/null 2>&1
	done
	/bin/rm -f /var/etc/nameserver_$1
fi
# Do not remove gateway used during filter reload.
/bin/rm -f /tmp/$1_router
/bin/rm -f /tmp/$1up
/bin/rm -f /tmp/$1_ip
/usr/local/sbin/pfSctl -c 'service reload dns'
ifconfig em0 down
ifconfig em0 up

Do not forget to replace em0 with your own device name.