WAN connection eventually dies after couple of days

cmb

Take the MAC out, then reboot to make sure. Not sure why you're seeing it that frequently, I ran that way for almost 2 years on 2.0 and only hit it 3 times, but it was eventually annoying enough I just took it off my DHCP WAN. But I'm sure that'll fix it, unless there's some other reason you start losing link on that NIC.

JohnnyBeGood

@cmb:

Take the MAC out, then reboot to make sure. Not sure why you're seeing it that frequently, I ran that way for almost 2 years on 2.0 and only hit it 3 times, but it was eventually annoying enough I just took it off my DHCP WAN. But I'm sure that'll fix it, unless there's some other reason you start losing link on that NIC.

I took the MAC out and rebooted modem + pfSense. Out of curiosity how did you figure out that spoofing was causing this issue? Any links to the ticket?

cmb

@JohnnyBeGood:

Out of curiosity how did you figure out that spoofing was causing this issue? Any links to the ticket?

I saw link cycling in your logs, and short of a physical cabling, modem or NIC issue that would cause that, which would probably be unlikely given a reboot makes it go away, it matches that issue.
http://redmine.pfsense.org/issues/1572

JohnnyBeGood

Thanks!

I replied to that post confirming same issue. Hopefully it gets fixed in the next release.
I'm glad that I have "stable" system not worrying when WAN connection will go down :)

shadow

Huh. I wonder if this (http://forum.pfsense.org/index.php/topic,36643.0.html) is this. Because, well, I am spoofing. I'll tweak and check it out.

JohnnyBeGood

So, far it's been running great and WAN its not dying like before. This is definitely a bug. Turning off mac spoofing fixed my problem.

Thanks guys!

shadow

Turned off spoofing (i was spoofing my real mac for some reason; i don't remember enabling this, but…) and now all is well. no more ~daily dropping offline on the wan.

GoldServe

Can anyone figure out what is causing the link to cycle when mac spoofing? I kind of need mac spoofing and this is a serious issue that requires a reboot!

jimp

There is an open ticket for it. I'm not sure if anyone has been able to track it down yet.

Most people only need MAC spoofing for a short-term fix (ARP cache issues, upstream equipment, etc) and once the situation resolves itself, they remove the spoofing.

tha_toadman

This issue has been a problem for me as of late. It may have been there when I upgraded to RC3 but only until recently have I noticed it.

I'm set for DHCP from my ISP with no MAC spoofing and every couple of days I'll randomly just get a DHCP failure on my boxes and my WAN connectivity drops out. I've had to release/renew to regain connectivity. I looked into the logs and I have an "apinger" entry that occurs around the same time as the DHCP trouble on my LAN. I failed to check tonight but typically this leads to a 169 IP (before the release/renew).


Aug 21 00:20:05 	apinger: alarm canceled: WAN(x.x.x.x) *** down ***
Aug 21 00:14:41 	check_reload_status: Reloading filter
Aug 21 00:14:31 	apinger: ALARM: WAN(x.x.x.x) *** down ***
Aug 20 21:49:55 	dhclient: Creating resolv.conf

FWIW, I saw another post mentioning the apinger issue and he was using the same modem that I am which is an SB6120. Perhaps there is an issue with this modem and pfSense? I can tell you that with this modem downstreams are bonded together for improved bandwidth. I'm not sure if that could be something of concern but I thought I would mention it.

Here is the post I was referring to…
http://forum.pfsense.org/index.php/topic,39205.msg202106.html#msg202106