Netgate Discussion Forum
    • Categories
    • Recent
    • Tags
    • Popular
    • Users
    • Search
    • Register
    • Login

    Cable Modem or Router Issue?

    Scheduled Pinned Locked Moved General pfSense Questions
    12 Posts 4 Posters 2.8k Views
    Loading More Posts
    • Oldest to Newest
    • Newest to Oldest
    • Most Votes
    Reply
    • Reply as topic
    Log in to reply
    This topic has been deleted. Only users with topic management privileges can see it.
    • B
      Bandalo
      last edited by

      I have a small home network, 3-4 PCs and a VM server.  I run pfsense in a VM with dedicated NICs for the LAN/WAN ports.  Haven't had any issues for years.  I noticed this morning I was seeing some odd behavior on my network.  Pages would not load, then after a couple refreshes, the would load just fine.  Then a few minutes later the problem repeats.  I checked the logs and saw this:

      Feb 5 06:20:45 apinger: ALARM: WAN(MY WAN IP) *** down ***
      Feb 5 06:20:55 check_reload_status: Reloading filter
      Feb 5 06:20:58 apinger: alarm canceled: WAN(MY WAN IP) *** down ***
      Feb 5 06:20:58 apinger: alarm canceled: WAN(MY WAN IP) *** loss ***
      Feb 5 06:21:06 apinger: alarm canceled: WAN(MY WAN IP) *** delay ***
      Feb 5 06:21:08 check_reload_status: Reloading filter
      Feb 5 06:21:24 apinger: ALARM: WAN(MY WAN IP) *** delay ***
      Feb 5 06:21:34 check_reload_status: Reloading filter
      Feb 5 06:21:44 apinger: ALARM: WAN(MY WAN IP) *** down ***
      Feb 5 06:21:52 apinger: alarm canceled: WAN(MY WAN IP) *** down ***
      Feb 5 06:21:54 check_reload_status: Reloading filter
      Feb 5 06:22:05 apinger: ALARM: WAN(MY WAN IP) *** down ***
      Feb 5 06:22:15 check_reload_status: Reloading filter
      Feb 5 06:22:17 apinger: alarm canceled: WAN(MY WAN IP) *** down ***
      Feb 5 06:22:25 apinger: alarm canceled: WAN(MY WAN IP) *** delay ***
      Feb 5 06:22:27 check_reload_status: Reloading filter
      Feb 5 06:22:51 apinger: ALARM: WAN(MY WAN IP) *** delay ***

      Now my cable modem shows a solid connection.  I had a similar problem last week and called the cable company (Comcast).  By the time they arrived the next day though, the problem had gone away.  They tested the house wiring, and the connections in the street, and didn't find any problems.  Any ideas of what is causing this problem?

      Thanks.

      1 Reply Last reply Reply Quote 0
      • S
        seraph8672
        last edited by

        I have this exact same issue with the exact same log messages. I was hunting the forum and found similar sounding issues such as this: http://forum.pfsense.org/index.php/topic,58240.0.html

        But I have done some additional testing that may prove useful:

        Originally I had been using pfsense since version 1 on this same hardware: a Dell Power Edge 810 with a P4 1GB of RAM and 2 onboard 10/100/1000 NICs
        it worked flawlessly for years.

        The other day I upgraded to 2.1Beta and suddenly I begin to see problems with packet loss on the WAN gateway. I swapped nics, I swapped cables, I  added a new NIC. No luck no matter which interface I assigned to WAN I would see packet loss alongside a corresponding CPU spike. So I tried a few things.

        To eliminate the possibility of a bad modem I connected my PC directly to the modem… no packet loss after 15 consecutive minutes of pinging 8.8.8.8 (google DNS server)

        Then I wanted to rule out my switches so I introduced them: Same result, flawless connection...

        I then took the gateway to my friends house so we could poke at the Ethernet drivers, but when I connected it to his equipment and routed my laptop through it.... NO ISSUES?!?! BWAH??!!

        I knew it was not my home connection, so I reasoned that it must be because I was already behind his gateway...  SO I took it home, and blew the dust off an old netgear router in a box to test it.

        Sure enough when I place the netgear router as the edge, and put the pfsense gateway behind it suddenly all the issues go away. Now I do not wish to keep this configuration as it throttles my connection at the WAP to 100Mb but it seems clear to me that there is some interaction between either my modem and pfsense, or my ISP and pfsense that is causing mayhem in the new version.

        tl;dr:
        my modem is good
        my connection is good
        this issue corresponds with a CPU spike
        packet loss only occurs when WAN interface is connected directly to the Modem

        Help? :)

        1 Reply Last reply Reply Quote 0
        • S
          sinicide
          last edited by

          Yeah it's such an annoying issue. For me as a way to alleviate some of this, I'm using this script.

          http://forum.pfsense.org/index.php/topic,17243.msg89348.html

          It auto-renews the connection after it sense it's been down. The Cronjob runs every three minutes. However the wan issue will continue to persist. I'm gonna get intel nics to try for my pfsense sense they're currently on realtek. I never use to have this issue on 2.0, but then I decided to upgrade….

          1 Reply Last reply Reply Quote 0
          • B
            Bandalo
            last edited by

            Well, I still can't really tell if it is definitely a router/pfsense issue, or something with the cable modem.

            I plugged my old Linksys router/AP in last night and seemed to see the same problem with connectivity, but it doesn't have any logs to speak of and no way to really diagnose the problem.  It does make me think that it's either a Comcast problem, or a cable modem problem.  So if you know any way to troubleshoot who's at fault there (besides buying a new cable modem) let me know!

            1 Reply Last reply Reply Quote 0
            • S
              seraph8672
              last edited by

              it makes no difference for me with intel NIC's as that is what I am using. Just for giggles to make sure it was certainly not my hardware, I used the live cd for zeroshell (I know, I know, I hate it too) and I have not lost a single packet through continuous testing over the last 8 hours. =/ I hate the zeroshell UI, and the stupid "Sorry we don't have an installer just use the live boot" crap, can we please get some love from pfsense?

              1 Reply Last reply Reply Quote 0
              • C
                cmb
                last edited by

                This may be the case of modems causing NIC link to cycle rapidly, then our check_reload_status skipped doing anything with the NIC to prevent other problems. This was fixed very recently in 2.0.3 and 2.1.

                1 Reply Last reply Reply Quote 0
                • S
                  seraph8672
                  last edited by

                  @cmb:

                  This may be the case of modems causing NIC link to cycle rapidly, then our check_reload_status skipped doing anything with the NIC to prevent other problems. This was fixed very recently in 2.0.3 and 2.1.

                  I just tried with last nights build of 2.1beta because I read it may have a fix for https://redmine.pfsense.org/issues/2792 which another thread suggested was the cause.  No luck

                  1 Reply Last reply Reply Quote 0
                  • B
                    Bandalo
                    last edited by

                    Would it matter in this case that pfSense is running in a VM?  It has a 2 port Intel NIC dedicated to it, one NIC port on a virtual switch dedicated to the WAN, and another dedicated to the LAN.

                    1 Reply Last reply Reply Quote 0
                    • C
                      cmb
                      last edited by

                      @Bandalo:

                      Would it matter in this case that pfSense is running in a VM?

                      Depends on the hypervisor. If it cycles the virtual NIC's link when the physical NIC cycles link, then yes. Otherwise no.

                      1 Reply Last reply Reply Quote 0
                      • B
                        Bandalo
                        last edited by

                        I'm running ESXI 5.1.0 using the default drivers for the Intel card..

                        I'm mainly trying to find some way to prove the problem is with the cable modem or connection rather than the router.

                        1 Reply Last reply Reply Quote 0
                        • S
                          seraph8672
                          last edited by

                          Well I'm using a physical machine for the gateway, my Pfsense VM that I use to DMZ my VM's is on intel nics in ESXi 5.0u4 and doesn't have this problem. I do see the link cycle on the WAN side of the phys gateway though… which is linked directly to the modem for WAN. Now my question to you is.... Do you have a motorola modem? Mine is the squarish Docsis 3.0 Motorola Modem with VOIP. I think it is handling DHCP req's by cycling the link fast instead of re broadcasting. Is this common? Is that how its supposed to work? idk. I only got through CCNA 1 and that was before cable modems were a thing >_< Vmware and big Storage are all I know.

                          1 Reply Last reply Reply Quote 0
                          • B
                            Bandalo
                            last edited by

                            I do have a Motorola cable modem, the 6121 I believe.

                            I haven't seen any unusual activity on the ESXI machine, but maybe I can dig through the network logs some more.

                            The problem has been intermittent..it went back to normal about a day after I posted this and has been working fine ever since.

                            1 Reply Last reply Reply Quote 0
                            • First post
                              Last post
                            Copyright 2025 Rubicon Communications LLC (Netgate). All rights reserved.