Connection lost for about 30seconds every 5 minutes approx. Pfsense or ISP?



  • I've been using pfSense for a long time now, and upgraded to 1.2 about a month ago. Yesterday i suddenly i started to get problems with the connection to internet.

    Every 5 minutes (approximately) i would loose connection for 10-60 seconds (approx). It kept going like this the whole night. In the morning i restarted the pfsense server, and then it apperaed to work OK! But then again, now the same problem appears! I tried restarting the server again, but this time it only worked for about 10 minutes before it happened again.

    This is what it says in the system logs:

    Mar 25 21:05:07 last message repeated 7 times
    Mar 25 21:02:32 miniupnpd[949]: Failed to get ip address for interface le1
    Mar 25 21:02:32 miniupnpd[949]: ioctl(s, SIOCGIFADDR, …): Can't assign requested address
    Mar 25 21:02:32 miniupnpd[949]: Failed to get ip address for interface le1
    Mar 25 21:02:32 miniupnpd[949]: ioctl(s, SIOCGIFADDR, …): Can't assign requested address
    Mar 25 21:00:37 miniupnpd[949]: Failed to get ip address for interface le1
    Mar 25 21:00:37 miniupnpd[949]: ioctl(s, SIOCGIFADDR, …): Can't assign requested address
    Mar 25 20:39:49 miniupnpd[949]: Failed to get ip address for interface le1
    Mar 25 20:39:49 miniupnpd[949]: ioctl(s, SIOCGIFADDR, …): Can't assign requested address
    Mar 25 20:39:49 miniupnpd[949]: Failed to get ip address for interface le1
    Mar 25 20:39:49 miniupnpd[949]: ioctl(s, SIOCGIFADDR, …): Can't assign requested address

    And check the below screenshot - every time the connection is lost, there's a spike… Only about 2mbps though - line is 12mbps so its not maxing it.

    Does anyone know how i can confirm whether or not this is ISP or pfsense's problem? And in case of pfsense - how i can investigate it?

    I was thinking of trying a downgrade to the previous firmware (some 1.2 beta), can someone just confirm that it's not a problem to downgrade?

    BTW: Only packages i have installed is bandwidthd and miniupnpd. Btw - under services - are these packages supposed to be listed twice? bandwidthd is listed twice with same description text, but minupnpd has two different descriptions.



  • What kind of WAN do you have? Is it DHCP? Maybe you are losing your IP-Adress on the WAN interface? If it was working for that long without issues and suddenly started without changing anything at your setup I would blame the ISP. However you have better chances to file a ticket if you can provide some detailed information on the issue of course.



  • pfSense checks the monitoring IP addresses every 5 minutes.  I would say that is your problem (invalid monitoring ip, etc).



  • sullrich: I'm not sure what you mean by "monitoring ip" - you mean its a ISP problem? I'm not loosing connection every 5 minutes - it appears to be very random, but in average i would guess 5 minutes.

    hoba: WAN is ADSL 2+ using PPPoE (static IP). WAN adapter is le1.

    The main reason why i'm thinking it may be pfSense is because it appeared to work fine for about 12 hours right after i restarted pfSense the first time. But of course, that could be a coincidence… But i also think those spikes right before connection is dropped was a little weird.

    Any other things i should check out before i contact my ISP?

    Edit: I'm just using  "ping -t" from Windows to monitor the connection - and the last 30 minutes it looks like instead of connection going down, the ping time just goes up to 100-200ms for the same period.. Only went completely down one time ("request timed out").
    Edit2: Now it just started to show "Request timed out" again... And it went down every 10-20 seconds the past minutes...



  • Woops, I am confusing two threads.  Try disabling UPNP and see if the problem persists.



  • @sullrich:

    Woops, I am confusing two threads.   Try disabling UPNP and see if the problem persists.

    I uninstalled the UPNP (and bandwidthd) package already, but it didn't make any difference. Although, the miniupnpd service is still listed under Services as stopped - and if i click start it says it was started, but still shows as "Stopped". Guess that's just a "visual" problem though.



  • UPNP is integrated in base in 1.2. You should not uninstall it or use the package anymore at all.



  • @hoba:

    UPNP is integrated in base in 1.2. You should not uninstall it or use the package anymore at all.

    I didn't install the package - i guess it was just keeping it after i upgraded to 1.2.. And now when i uninstalled it, all UPNP links in the menu are no longer working. The package is not listed among the available packages for install either…

    Would it be smart to reinstall the entire server, and restore settings from backup? Or would that be pretty much the same as an upgrade?



  • I would reinstall to rule out any existing updateglitch that you might have on your system atm (and something seems to be obviously wrong).



  • @GeeZuZz:

    sullrich: I'm not sure what you mean by "monitoring ip" - you mean its a ISP problem? I'm not loosing connection every 5 minutes - it appears to be very random, but in average i would guess 5 minutes.

    hoba: WAN is ADSL 2+ using PPPoE (static IP). WAN adapter is le1.

    The main reason why i'm thinking it may be pfSense is because it appeared to work fine for about 12 hours right after i restarted pfSense the first time. But of course, that could be a coincidence… But i also think those spikes right before connection is dropped was a little weird.

    Any other things i should check out before i contact my ISP?

    Edit: I'm just using  "ping -t" from Windows to monitor the connection - and the last 30 minutes it looks like instead of connection going down, the ping time just goes up to 100-200ms for the same period.. Only went completely down one time ("request timed out").
    Edit2: Now it just started to show "Request timed out" again... And it went down every 10-20 seconds the past minutes...

    If you think it is the ISP, you can try this…..

    open 5 cmd windows on your windows box
    first cmd window ping your internal gateway EX: 192.168.1.1 -t
    second cmd window ping your WAN port Static ip EX: 10.1.1.10 -t
    At this point…you **"should"** not have any drops if it is your ISP
    now the 3rd cmd window ping your ISP gateway IP 10.1.1.1 -t
    the 4th, just ping something on the net EX 4.2.2.2 -t
    5th ping another internet address EX ping 208.68.222.222 -t

    run these pings for a while to monitor
    If you can, isolate one pc on the LAN to do the testing.
    If you think it is PFSense, take the box out of the picture by connecting your pc/laptop right to the provider device (with your windows firewll on ;) ) and setup your connection to the net…run the same test

    Now, if you get drops past the 3 cmd window (ISP gateway) , good chances are that your ISP is the problem.
    If you get drops on 3,4 and 5 all at the same times, check with your ISP.

    You can also use something link ping plotter to ping these address over a timeframe and you will be able to check what time they dropped.


Locked