Is this a bug on Asus firmware? WAN disconnections

Discussion in 'Asuswrt-Merlin' started by guipel, Oct 30, 2012.

  1. guipel

    guipel Occasional Visitor

    Joined:
    Aug 13, 2012
    Messages:
    29
    Location:
    Brazil
    Hello guys,

    Maybe I found a bug on Asus firmware.

    I'm having some WAN disconnection issues, and first I suspected about my cable ISP service.

    When I lose the connection I am able to ping my router (192.168.1.1), even my cable modem (192.168.100.1), but I am not able to ping any external server like 8.8.8.8 (Google DNS)

    Initially I was solving this justing restarting my cable modem. But this is very annoying, because happens a lot of times per day.

    So last time this problem occurred I reseted the log on router and pulled the WAN cable off:

    When I reconnected the WAN cable:

    So I tried changing to my another router (an old Apple Airport Express) and this problem never happened again.

    Tried back on Asus and the issue returned.

    I am trying to download the source code to look for something but Asus is very slow at this moment.

    Any thoughts?

    Best
     
  2. Log in / Register to remove this ad

  3. RMerlin

    RMerlin Part of the Furniture

    Joined:
    Apr 14, 2012
    Messages:
    14,736
    Location:
    Canada
    I bet you are using a Cat5 Ethernet cable between the router and the modem. RT-N66U has a gigabit WAN interface, so it needs to be either Cat5e or Cat6.

    Also when switching a cablemodem to a new router, you need to fully reset the connection, usually by leaving the cablemodem turned off for 5-10 mins.
     
  4. guipel

    guipel Occasional Visitor

    Joined:
    Aug 13, 2012
    Messages:
    29
    Location:
    Brazil
    Hello Merlin, appreciate your help.

    I double checked and I am using a cat5e.
    Will I have a better chance with cat6?

    When I switched the routers I reseted the modem. I tested for a few hours and have no issue. When I turned back to asus had the problem again in just a few minutes.

    When this problem is occurring all the lights on modem are OK, just like there isn't no problem.

    I already called my ISP, that changed the cable modem to a brand new one and the problem persists :(

    If I just try to reset the router nothing happens, I still have the problem. I am guessing in some issue regarding DHCP lease or something like that.
     
  5. guipel

    guipel Occasional Visitor

    Joined:
    Aug 13, 2012
    Messages:
    29
    Location:
    Brazil
    Right now everything is OK, next time that happens I will check the status of these two variables in nvram:

     
  6. RMerlin

    RMerlin Part of the Furniture

    Joined:
    Apr 14, 2012
    Messages:
    14,736
    Location:
    Canada
    No, Cat5e and Cat6 are both gigabit-capable, they would both work just as well in this setup.
     
  7. lollomanta

    lollomanta New Around Here

    Joined:
    Nov 3, 2012
    Messages:
    2
    I also have the same problem. After i put my modem in bridge, i loose my internet connection a few times each day. My log show this:

    Nov 1 00:31:24 WAN Connection: Ethernet link down.
    Nov 1 00:31:24 notify_rc : stop_nat_rules
    Oct 31 23:31:24 stop_nat_rules: apply the redirect_rules!
    Nov 1 00:31:47 WAN Connection: Ethernet link up.
    Nov 1 00:31:47 notify_rc : restart_wan_if 0
    Nov 1 00:31:47 dnsmasq[505]: read /etc/hosts - 3 addresses
    Nov 1 00:31:47 dnsmasq-dhcp[505]: read /etc/ethers - 2 addresses
    Nov 1 00:31:47 dnsmasq[505]: using nameserver 193.213.112.4#53
    Nov 1 00:31:47 dnsmasq[505]: using nameserver 130.67.15.198#53
    Oct 31 23:31:47 stop_wan(): perform DHCP release
    Nov 1 00:31:47 dnsmasq[505]: read /etc/hosts - 3 addresses
    Nov 1 00:31:47 dnsmasq-dhcp[505]: read /etc/ethers - 2 addresses
    Nov 1 00:31:47 dnsmasq[505]: using nameserver 193.213.112.4#53
    Nov 1 00:31:47 dnsmasq[505]: using nameserver 130.67.15.198#53
    Nov 1 00:31:47 miniupnpd[1266]: ioctl(s, SIOCGIFADDR, ...): Cannot assign requested address
    Nov 1 00:31:47 miniupnpd[1266]: Failed to get ip address for interface eth0
    Nov 1 00:31:47 kernel: br0: port 1(vlan1) entering disabled state
    Nov 1 00:31:47 kernel: br0: port 1(vlan1) entering listening state
    Nov 1 00:31:47 kernel: br0: port 1(vlan1) entering learning state
    Nov 1 00:31:47 kernel: br0: topology change detected, propagating
    Nov 1 00:31:47 kernel: br0: port 1(vlan1) entering forwarding state
    Nov 1 00:31:49 dnsmasq[505]: read /etc/hosts - 3 addresses
    Nov 1 00:31:49 dnsmasq-dhcp[505]: read /etc/ethers - 2 addresses
    Nov 1 00:31:55 WAN Connection: ISP's DHCP did not function properly.

    I think it happens when the ip-lease expires
     
  8. xrecar

    xrecar Regular Contributor

    Joined:
    Jun 14, 2012
    Messages:
    64
    The only way I have been able to fix my disconnect is to reinstall merlins firmware. After the reinstall the disconnected returns to connected. Happens every 2 to 3 days...

    I've posted this issue a few times, never had an answer yet.:confused:

    Good idea about the lease, I have 5 hrs left and will pay attention.
     
    Last edited: Nov 3, 2012
  9. guipel

    guipel Occasional Visitor

    Joined:
    Aug 13, 2012
    Messages:
    29
    Location:
    Brazil
    After a lot of research here, this problem is gone.

    I made a lot of tests and the last time I switched to my Apple Airport Express and returned to Asus, my IP Address changed from 189.x.x.x to 201.x.x.x

    I guess the change in MAC address forced the ISP to put me on another DHCP server. After this change the problem is gone.

    Unfortunately I don'd now what was causing the issue when I had the IP 189.x.x.x

    Hope this helps.

    Best
     
  10. octopus

    octopus Senior Member

    Joined:
    Jul 17, 2012
    Messages:
    373
    Some ISP send DHCP renew/relese on port 67 and 68 so adding line in iptable is needed:
    That can cause dissconnect when renew is sent.

    octopus
     
    Last edited: Nov 4, 2012
  11. lollomanta

    lollomanta New Around Here

    Joined:
    Nov 3, 2012
    Messages:
    2
    Where do i find that iptable?

    I have one hour leasetime. I loose connection a few times each day. According to the log, it only happened 1 time during the last 12 hours. After a minute it's up running again
     
  12. octopus

    octopus Senior Member

    Joined:
    Jul 17, 2012
    Messages:
    373
    I don't know where you can put it in, I'm usually using Tomato but that was problem long time ago.
    This not affect your LAN lease time, lease time from your ISP. I have 4 hour from my ISP.
     
    Last edited: Nov 4, 2012
  13. RMerlin

    RMerlin Part of the Furniture

    Joined:
    Apr 14, 2012
    Messages:
    14,736
    Location:
    Canada
    That rule is already included in the default set of iptable rules.
     
  14. mdusen1216

    mdusen1216 New Around Here

    Joined:
    Nov 11, 2012
    Messages:
    2
    I am experiencing this same issue.

    Dec 31 19:05:13 WAN Connection: Ethernet link down.
    Dec 31 19:05:14 WAN Connection: Ethernet link up.
    Dec 31 19:05:14 notify_rc : restart_wan_if 0
    Dec 31 19:05:14 dnsmasq[517]: read /etc/hosts - 3 addresses
    Dec 31 19:05:14 dnsmasq-dhcp[517]: read /etc/ethers - 7 addresses
    Dec 31 19:05:14 stop_wan(): perform DHCP release
    Dec 31 19:05:14 kernel: br0: port 1(vlan1) entering disabled state
    Dec 31 19:05:15 kernel: br0: port 1(vlan1) entering listening state
    Dec 31 19:05:15 kernel: br0: port 1(vlan1) entering learning state
    Dec 31 19:05:15 kernel: br0: topology change detected, propagating
    Dec 31 19:05:15 kernel: br0: port 1(vlan1) entering forwarding state
    Dec 31 19:05:17 dnsmasq[517]: read /etc/hosts - 3 addresses
    Dec 31 19:05:17 dnsmasq-dhcp[517]: read /etc/ethers - 7 addresses
    Dec 31 19:05:17 WAN Connection: ISP's DHCP did not function properly.


    The only way I am able to bring the Wan back up is to change the WANs Mac address then power cycle the Modem. So far it has happened 2 days in a row now at almost the exact same time. The modem (Motorola sb6121) and router Asus RT-N66R by the way are brand new.
     
  15. RMerlin

    RMerlin Part of the Furniture

    Joined:
    Apr 14, 2012
    Messages:
    14,736
    Location:
    Canada
    Try leaving your modem off for 5-10 minutes, to ensure it gets fully reset at your ISP's end. Some cablemodem ISPs can be picky with allocating WAN leases.
     
  16. NeutralGold

    NeutralGold Regular Contributor

    Joined:
    May 2, 2012
    Messages:
    123
    Very picky indeed. I use comcast and when I fully reset my modem, I get a 1 hour lease. Once that's up I get something like a 5 hour lease and another at 1 day until finally I get 3 days 20 hours 7 minutes 11 seconds. Several times I've also been placed on completely different subnets. I experience this every firmware upgrade/swap.
     
  17. mdusen1216

    mdusen1216 New Around Here

    Joined:
    Nov 11, 2012
    Messages:
    2
    Thanks RMerlin but that is definately not what is happening here. This router will simply drop its WAN dhcp settings and be unable to renew them unless I do the things I have described in my previous post. If it is simply renewing the dhcp lease from my ISP the modem should not require a power reset unless the hardware address of the router has changed.
     
  18. xrecar

    xrecar Regular Contributor

    Joined:
    Jun 14, 2012
    Messages:
    64
    What happens when you are disconnected and reinstall the firmware? Mine magically reconnects. Its like something in the fw resets on reinstall/reboot. Nothing happens on a plain reboot. Funny I have the same cable modem sb6121.

    This just happened;

    Nov 16 02:24:01 WAN Connection: WAN was restored.
    Nov 16 02:24:01 notify_rc : start_nat_rules
    Nov 16 02:24:28 dhcp client: bound 192.168.100.10 via 192.168.100.1
    Nov 16 02:24:46 dnsmasq[513]: read /etc/hosts - 3 addresses
    Nov 16 02:24:46 dnsmasq[513]: using nameserver 192.168.100.1#53
    Nov 16 02:24:46 WAN Connection: ISP's DHCP did not function properly.

    When I check the Mac address, it has changed from what it was cloned from originally? What would cause that?

    Do you think it could be a modem issue?
     
    Last edited: Nov 15, 2012
  19. SoCalReviews

    SoCalReviews Regular Contributor

    Joined:
    Aug 16, 2012
    Messages:
    149
    I was having similar issues with disconnects from my SB6121 cable modem a few months ago and called support. I hadn't had any problems for over a year and then this year in early August 2012 it seemed I began getting disconnects and the WAN IP wouldn't update. I can't say whether these disconnects had anything at all to do with the Asus router or firmware updates because I hadn't had problems with the RT-N66U router and the cable internet for the month or two prior to these incidents. The cable ISP support told me that if unplugging the power from the modem for 5-10 minutes didn't work that I should also try physically unscrewing the cable from the SB6121 cable modem as well as unplugging power for 5-10 minutes.

    The support person didn't have a logical explanation as to why also physically disconnecting the cable worked for so many cable customers who couldn't renew their connection but I would theorize that the cable company equipment may sense the physical connection even when the cable modem power is off and sometimes they may not renew the WAN IP or maybe the modem holds onto the old WAN IP and won't renew even when the power is off. This is the only logical theory I can come up with why physically removing the cable would work.

    I have done this twice to successfully correct the problem of loss of connection and unable to renew when only powering the modem off for 5-10 min didn't seem to work. One of those times cable company support told me that they had been working on the network in my area when my internet was down for an hour. The other time I didn't call support and I just did the same dual procedure of power off and physically disconnect the cable from the modem for 5-10 min.. Then when I reconnected the cable and plugged the power back into my SB6121 it was connecting up and working perfectly. That was about a month ago and I haven't had any problems since then. I will have to wait and see if this problem happens again in the future.

    Another note of possible importance is that since July 2012 many U.S. cable companies have updated their ISP terms of agreement and they are now more closely monitoring and recording customer internet access...sites accessed by customers, equipment, etc.. There may be underlying U.S. internet laws requiring larger ISPs to monitor their customers access and usage. Many larger ISPs are moving away from allowing customer owned modems and equipment and requiring the use of their modems. Anyone who is using the newer SB6141, SB6121 or other personally owned cable modem should make sure that their cable company has properly provisioned it and included their latest updates for that equipment.
     
    Last edited: Nov 28, 2012
  20. petrmegac

    petrmegac New Around Here

    Joined:
    Jan 6, 2013
    Messages:
    4
    Hi, I have the same problem.

    Jan 1 01:00:43 ddns update: no update needed at this time
    Jan 1 01:00:43 WAN Connection: WAN was restored.
    Jan 1 01:00:43 dnsmasq[510]: read /etc/hosts - 3 addresses
    Jan 1 01:00:43 dnsmasq-dhcp[510]: read /etc/ethers - 7 addresses
    Jan 1 01:00:43 dnsmasq[510]: using nameserver 93.153.117.1#53
    Jan 1 01:00:43 dnsmasq[510]: using nameserver 93.153.117.33#53
    Jan 1 01:00:43 dnsmasq[510]: using nameserver 195.113.144.194#53
    Jan 1 01:00:43 dnsmasq[510]: using nameserver 8.8.4.4#53
    Jan 1 01:00:43 dhcp client: bound 78.45.228.2 via 78.45.228.1
    Jan 1 01:00:48 WAN Connection: ISP's DHCP did not function properly.
    Jan 1 01:00:48 notify_rc : stop_nat_rules
    Jan 1 01:00:48 stop_nat_rules: apply the redirect_rules!

    This problem occurs when I register DDNS asuscomm.com. When I turn on the cable modem CISCO and router RT-AC66U (firmware 3.0.0.4.266.23b) the internet is not.
    Then I have to manually restart the router and everything is OK.

    If I cancel the DDNS service, everything is fine after switching ON cable modem and router.

    I think it's a firmware bug but I'm not an expert. Thank you for your advice. Sorry for the bad English.
     
  21. sovereignty68

    sovereignty68 New Around Here

    Joined:
    Apr 30, 2012
    Messages:
    8
    I also have this same problem with RT-N66U, RT-AC66U and Mikrotik OS Routerboard. They share exact same problem. They were unable to obtain/renew IP address from the modem (SB6121, SB6141, and uBee) once it was released. Rebooting the router doesn't work, reboot modem doesn't work either. Resetting them to defaults or changing MAC address of WAN interface didn't work either. I tried MK's forum but no one was able to find out what was the problem. From the log, all I can see is Modem (ISP) is't issue the IP. The most effective workaround for me to fix this problem is backup the settings, reflash the firmware (in recovery mode for Asus routers or system upgrade for Mikrotik system).
     
    Last edited: Jan 10, 2013

Share This Page