WAN looses connection randomly with 24-36 hours - tried everything
-
Yup. And that can be some upstream problem. Cable degraded causing the modem to lose sync for example.
-
Called my ISP, they're looking into the issue as well. Will report back, when if something happens again.
-
Now there's new stuff in the logs:
Jul 6 15:12:34 kernel arpresolve: can't allocate llinfo for 2.106.185.193 on igb0 Jul 6 15:12:34 kernel arpresolve: can't allocate llinfo for 2.106.185.193 on igb0 Jul 6 15:12:34 kernel arpresolve: can't allocate llinfo for 2.106.185.193 on igb0 Jul 6 15:12:34 kernel arpresolve: can't allocate llinfo for 2.106.185.193 on igb0 Jul 6 15:12:34 kernel arpresolve: can't allocate llinfo for 2.106.185.193 on igb0 Jul 6 15:12:34 kernel arpresolve: can't allocate llinfo for 2.106.185.193 on igb0 Jul 6 15:12:34 check_reload_status rc.newwanip starting igb0 Jul 6 15:12:34 check_reload_status Restarting ipsec tunnels Jul 6 15:12:35 php-fpm 78624 /rc.newwanip: rc.newwanip: Info: starting on igb0. Jul 6 15:12:35 php-fpm 78624 /rc.newwanip: rc.newwanip: on (IP address: 2.106.185.197) (interface: WAN[wan]) (real interface: igb0). Jul 6 15:12:35 dhcpleases /etc/hosts changed size from original! Jul 6 15:12:35 dhcpleases Could not deliver signal HUP to process because its pidfile (/var/run/unbound.pid) does not exist, No such process. Jul 6 15:12:36 dhcpleases /etc/hosts changed size from original! Jul 6 15:12:36 php-fpm 78624 /rc.newwanip: Removing static route for monitor 8.8.8.8 and adding a new route through 2.106.185.193 Jul 6 15:12:36 dhcpleases Could not deliver signal HUP to process because its pidfile (/var/run/unbound.pid) does not exist, No such process. Jul 6 15:12:37 dhcpleases kqueue error: unknown Jul 6 15:12:38 check_reload_status updating dyndns wan Jul 6 15:12:39 php-fpm 27344 /rc.dyndns.update: phpDynDNS (): No change in my IP address and/or 25 days has not passed. Not updating dynamic DNS entry. Jul 6 15:12:40 dhcpleases /etc/hosts changed size from original! Jul 6 15:12:40 php-fpm 67720 /interfaces.php: Removing static route for monitor 8.8.8.8 and adding a new route through 2.106.185.193 Jul 6 15:12:40 check_reload_status Reloading filter Jul 6 15:12:40 php-fpm 67720 /interfaces.php: Creating rrd update script Jul 6 15:12:40 snmpd 84244 disk_OS_get_disks: adding device 'ada0' to device list Jul 6 15:12:41 php-fpm 78624 /rc.newwanip: The command '/usr/local/sbin/unbound -c /var/unbound/unbound.conf' returned exit code '1', the output was '[1594041161] unbound[22452:0] error: bind: address already in use [1594041161] unbound[22452:0] fatal error: could not open ports'
-
@jim82 said in WAN looses connection randomly with 24-36 hours - tried everything:
Called my ISP, they're looking into the issue as well. Will report back, when if something happens again.
@stephenw10 is 100% spot on. Likely upstream issue with your provider.
When you have a cable modem, your pfSense WAN is configured to use DHCP to obtain its IP address. The cable modem is the provider of that DHCP address. The cable modem operates in two modes. The normal mode, when it is properly synchronized with the CMTS (cable modem termination system) equipment of your cable ISP, is for the cable modem to pass the IP it gets from the CMTS on to pfSense using DHCP. However, when an issue occurs out in the cable TV coax plant such that your modem loses sync with the CMTS, then the cable modem goes into "private DHCP mode" and issues attached devices a local non-routable IP address (typically in the 192.168.100.0/24 subnet).
So long as pfSense sees an IP address on its WAN, it thinks things are fine and tries to use it. But that 192.168.100.x IP is not good for the internet, so nothing works properly. Ideally, when the modem loses sync and goes into that private DHCP mode, it should tell pfSense when it gets back in touch with the CMTS and obtains a proper IP address and let pfSense restart the WAN interface with the new address. Unfortunately that can sometimes not happen properly and pfSense winds up stuck with the non-routable private address.
Telling pfSense to ignore DHCP offers from the cable modem's private DHCP pool (that 192.168.100.0/24 pool) can prevent pfSense from getting "stuck" on the private non-routable IP. That still won't stop your upstream issue that is the root cause, but it will prevent pfSense from getting stuck on one of those non-routable IP addresses whenever loss of sync happens.
-
Du kører med en Yousee forbindelse.
Sagem modem skal i bridge mode og være helt transparent. Det klarer yousee support.
Dernæst konfigurerer du DHCP på pfsense og lader den stå der.
Så kører det.
-
@bmeeks Thanks a lot for the indepth explanation. That sure makes sense. Hoping my ISP is able to see why it's loosing sync.
-
Current status:
Talked to a technician from my ISP today. He reports that Sagem apparently have issues with intermittent "freeze ups" when OFDN channels are activated on upstream for the Docsis 3.1 standard.
For now, he has deactivated those channels on my connection, which apparently removes the "freeze" problem. Fingers crossed, that it will also remove my sync problem.
New firmware is being worked on.
-
Huh, that sounds fun. But that's a good response from an ISP, most would never admit any fault exists.
Steve
-
So far, so good. It's been 48+ hours since the last dropout. Looks like the removal of OFDM** upstream channels, solved the issue. Now I'm waiting for the new firmware to arrive.
Not a single error in the logs since. Thanks for your help so far.
-
14 days after and my connection is rock solid. Thanks for your help.
-
@jim82 said in WAN looses connection randomly with 24-36 hours - tried everything:
14 days after and my connection is rock solid. Thanks for your help.