Gateway packet loss % incorrect



  • When receiving lots of traffic on a WAN interface, the gateway for that interface is marked as high packet loss with a loss percentage higher than 100%!

    Also, when manually pinging that same gateway from the console there is 0% packet loss. I made sure to ping out the correct interface.

    11 packets transmitted, 11 packets received, 0.0% packet loss
    round-trip min/avg/max/stddev = 16.007/16.950/19.544/0.923 ms
    

    This is on:
    2.1-RC2 (amd64)
    built on Thu Sep 5 21:38:32 EDT 2013
    FreeBSD 8.3-RELEASE-p10



  • Well - I'm pretty sure the only way to lose 192% of anything is in a divorce, so I don't think the packet loss number can be correct?



  • Is this with a recent snapshot?
    There were issues a week or 2 ago with some apinger changes.



  • I have modified the original post to contain the build this is happening on.



  • On one of my pfSense sites, it occasionally gets very high pings upwards of 900ms.

    I've adjusted apinger's monitor IP thresholds to some ridiculous numbers and I still get WAN delay and WAN loss warnings causing my OpenVPN connection to flap.  I wonder if the OP's observations are a cause for the symptoms I've been seeing with apinger.

    The pings at that site are high with some minimal packet loss but nowhere near the amount to set off the thresholds I've set below.

    Sept. 6 Snapshot

    ![Screen Shot 2013-09-08 at 2.31.17 PM.png](/public/imported_attachments/1/Screen Shot 2013-09-08 at 2.31.17 PM.png)
    ![Screen Shot 2013-09-08 at 2.31.17 PM.png_thumb](/public/imported_attachments/1/Screen Shot 2013-09-08 at 2.31.17 PM.png_thumb)



  • I just checked now and, even though there is no more traffic on that particular WAN interface, packet loss is still showing 192%, the same value as in the OP. The "Last check" field shows my correct local time. Is it possible for the packet loss percentage to get stuck on such a weird value?



  • Math can be weird at times :).
    It will recover itself after some successful pings and the same issue has been there since 2.0.x when apinger got introduced just not so visible due to apinger getting killed and restarted and now only reloading the config.

    It will be correctly fixed for 2.1.x+ to avoid these oddities.
    Its just math issue not more from the value point of view.

    Is it triggering any alarms due to this?



  • This can't be good… ::)  Anybody figure anything out on this one?



  • Rebel Alliance Global Moderator

    And what version are you running - the same as the OP?

    2.1-RC2 (amd64)
    built on Thu Sep 5 21:38:32 EDT 2013

    Or the current build?



  • @johnpoz:

    And what version are you running - the same as the OP?

    2.1-RC2 (amd64)
    built on Thu Sep 5 21:38:32 EDT 2013

    Or the current build?

    2.1-RELEASE (i386)
    built on Wed Sep 11 18:16:50 EDT 2013
    FreeBSD 8.3-RELEASE-p11



  • i've got the same issue.

    Version:
    2.1-RELEASE (amd64)
    built on Wed Sep 11 18:17:48 EDT 2013
    FreeBSD 8.3-RELEASE-p11

    to fix i restart the apinger services, but it isn't a permanent fix, after some hour the issue come back.



  • Looks like I may have the same issue on 2.1.1; pinging from the command line I get no packet loss, but apinger/pfSense reports 2% packet loss or higher!



  • Are you pinging the same device as apinger? I regularly get packet-loss to my ISP's gateway because it has an ICMP limit. It doesn't actually reflect my connection's packet-loss.



  • I have the same behavior in pfSense 2.1.3, after restart the apinger service everything is fine.
    But it is not a permanent solution for me that i have to restart the apinger service serval times a month … :-(

    Currently, there are no permanent solution?



  • Seeing the same in pfSense 2.1.4 - packet loss was showing 179% and restarting the apinger service cleared it up



  • Just experienced a similar problem. Gateways had failed to OPT1 from WAN, WAN showing 36% packet loss. RRD showed packet loss happening for proceeding 7 hrs. Console checking showed no packet loss. Reset of apinger reset gateway monitor, failover, and rrd graphs from that point forward. Both WAN and WAN_Ipv6 showed same results. NOTE: Comcast buried the cable at least 8 hrs earlier. It's possible that there was an interruption at some point, but I can't reproduce it.

    2.1.4-RELEASE(amd64) nanobsd(4g) running on AMD G-T40E.



  • I encountered consistent massive 308% packet loss on WAN gateway. Restarted cable modem first, re-plugged all cables in chain, just to be sure, but no avail. pfSense reboot cured it…
    Stumbled upon this thread afterwards... :( So for the statistics I'm adding to this thread, who knows it helps.
    Next time I'll try to restart the apinger service seperately.

    2.1.4-RELEASE(amd64) running on APU1C



  • There's a "Four Yorkshiremen" sketch in here somewhere…

    "308% packet loss??  LUXURY!!  Our router was a wet paper bag in the middle of the road!  We had to get up two hours before we went to bed, craft the packets by hand, send them uphill both ways..."



  • @KOM:

    "308% packet loss??  LUXURY!!

    I must admit I'm spoiled with my ISP for a home gateway… 160/8 isn't that bad  ;D  But shhh... don't tell them  ;)