Loss of TV signal after 4 hours (passthrough IP router)

  • Hello everyone,

    I have a centro piccolo configured in manual IP passthrough mode (since it is the only one now available), and a netgear r7000 flashed with DDWRT behind, which is configured with the swisscom WAN IP. The piccolo doesn’t do much, apart from DHCP for the TV2.0 box. Everything else is behind on the r7000.

    I attach an overview diagram.

    home_network.png

    Everything works, including TV2.0 which is anyway connected directly to the piccolo’s butt to avoid any QoS or IGMP inconvenience. Except that after 4 hours (and it really seems to be 4 hours quite reliably, still “measuring”), loss of channel signal (The Internet has no problem).

    What’s weird is that when changing the channel, the video reappears for around 30 secs, then loss of signal again, etc… Which would seem to point to a multicast problem, I think that the first 30 secs are maybe unicast?

    However, if I turn TV2.0 off and on again, nothing at all, not even for the first 30 seconds.

    The workaround to get signal back for a few hours is as follows:

    * Leave the TV 2.0 box on with its signal loss…

    * Turn off the piccolo router

    * Unplug the cable connecting the piccolo to the netgear r7000

    * Restart the piccolo router, wait until it is synchronized and has internet.

    --> The TV then comes back on.

    * Reconnect the cable between the piccolo and the netgear r7000

    --> Everything works, and the TV repeats 4 hours later.

    We could therefore assume that having the WAN IP shared thanks to IP passthrough poses a problem. But I haven’t been able to diagnose anything so far. And above all, why does it work for 4 hours, then suddenly nothing?

    Thank you all!

    ++

    Fred

    Show original language (French)
    • 6 p.m. without TV interruption (and no other problem), after putting the piccolo back on.

      The only thing I can see is the following, on the piccolo (overview of connected devices)

      nw_devices.jpg

      This r7000 interface is the secondary IP assigned to the r7000 WAN interface, which is also assigned the passthrough WAN IP.

      vlan2 Link encap:Ethernet HWaddr E4:F4:C6:18:19:EB
      inet addr:83.78.153.217 Bcast:83.78.153.255 Mask:255.255.255.0
      UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1
      RX packets:6364543 errors:0 dropped:10957 overruns:0 frame:0
      TX packets:4138096 errors:0 dropped:0 overruns:0 carrier:0
      collisions:0 txqueuelen:1000
      RX bytes:6815391908 (6.3 GiB) TX bytes:827960153 (789.6 MiB)

      vlan2:0 Link encap:Ethernet HWaddr E4:F4:C6:18:19:EB
      inet addr:192.168.1.254 Bcast:192.168.1.255 Mask:255.255.255.0
      UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1

      Previously, in this same overview of devices, the piccolo listed 2 “r7000” (same MAC address): the one shown above AND a duplicate with the WAN IP.

      Was this the cause of the problem? Was multicast trying to get the hell out of the r7000, while the picollo also had the WAN IP and therefore the interface locally, and logically with a higher priority? In any case, a tcpdump on the r7000 on the vlan2 interface did not reveal any multicast traffic so good.

      In any case, no problem since 6 p.m. now. It remains a mystery all the same.

    Follow-up

    After a call to swisscom support and a reset (and upgrade firmware) of my piccolo router, the IP passthrough function improved! I was able to choose dhcps-fixed, which was not the case before (function not available, only manual could be chosen).

    --> So no need for a script on my r7000 to update my WAN IP

    --> Maybe no more TV problem; I will keep this post updated

    #Centro_piccolo (ip wan-allocation)>> set
    wan-allocation
    mode (passthrough) [ normal | defaultserver | passthrough]:passthrough
    passthrough-mode (manual) [ manual | dhcps-fixed | dhcps-fcfs]:dhcps-fixed
    fixed-MAC-address (""): e4:f4:c6:18:19:eb
    lease-time (600) [ 180 - 8553600 ]:

    #Centro_piccolo (ip wan-allocation)>> save

    To be continued!

    THANKS,

    Fred

    Show original language (French)

    After trying the CentroGrande, it appears that the ADB firmware (latest version) does not react correctly to IP passthrough at all, at least not with the r7000.

    With equal configuration for the piccolo and the r7000:

    * the WAN IP goes well to the r7000

    * No communication possible between the r7000 and the CentroGrande.

    * With a secondary IP configured on the r7000 in the 192.168.1.0/24 network, it is now possible to ping Grande, 8.8.8.8 but not possible to ping google.com; DNS resolution does not work (even putting openDNS to see).

    So give up with the Grande. After reconnecting the Piccolo, and without any other configuration changes, it works directly (except that the TV will shut down within a few hours).

    After an interesting discussion with a level 2 at Swisscom, it could be that the fact that the TV2.0 box tries to go into standby after a few hours causes the problem. Unfortunately, I didn’t find any options to disable this mode. And then it remains to be seen why anyway.

    In fact, I asked for a standard Internet-Box in exchange for the Grande, and I will probably have to decide to give up using my r7000 since this new model does not support IP passthrough or other bridge mode.

    I imagine it’s my wife and my children who will be happy that I stopped “breaking the internet” 😉

    If others have feedback, I’ll keep the piccolo… 😉

    Thanks for following my monologue 🙂

    ++

    Fred

    Show original language (French)

    6 p.m. without TV interruption (and no other problem), after putting the piccolo back on.

    The only thing I can see is the following, on the piccolo (overview of connected devices)

    nw_devices.jpg

    This r7000 interface is the secondary IP assigned to the r7000 WAN interface, which is also assigned the passthrough WAN IP.

    vlan2 Link encap:Ethernet HWaddr E4:F4:C6:18:19:EB
    inet addr:83.78.153.217 Bcast:83.78.153.255 Mask:255.255.255.0
    UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1
    RX packets:6364543 errors:0 dropped:10957 overruns:0 frame:0
    TX packets:4138096 errors:0 dropped:0 overruns:0 carrier:0
    collisions:0 txqueuelen:1000
    RX bytes:6815391908 (6.3 GiB) TX bytes:827960153 (789.6 MiB)

    vlan2:0 Link encap:Ethernet HWaddr E4:F4:C6:18:19:EB
    inet addr:192.168.1.254 Bcast:192.168.1.255 Mask:255.255.255.0
    UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1

    Previously, in this same overview of devices, the piccolo listed 2 “r7000” (same MAC address): the one shown above AND a duplicate with the WAN IP.

    Was this the cause of the problem? Was multicast trying to get the hell out of the r7000, while the picollo also had the WAN IP and therefore the interface locally, and logically with a higher priority? In any case, a tcpdump on the r7000 on the vlan2 interface did not reveal any multicast traffic so good.

    In any case, no problem since 6 p.m. now. It remains a mystery all the same.

    Show original language (French)