Date: Fri, 20 May 2005 09:02:45 +0300 From: "Donatas" <donatas@lrtc.net> To: "Gleb Smirnoff" <glebius@FreeBSD.org> Cc: freebsd-net@freebsd.org Subject: Re: ng_bridge fluctuations Message-ID: <006101c55d01$8b08fe20$9f90a8c0@DONATAS> References: <131b01c55b9a$5fb58530$9f90a8c0@DONATAS> <20050518112720.GA6678@cell.sick.ru>
next in thread | previous in thread | raw e-mail | index | archive | help
> Donatas, > > On Wed, May 18, 2005 at 02:11:43PM +0300, Donatas wrote: > D> D> ### Regular MRTG monitoring > D> D> 0,10,20,30,40,50 * * * * root /usr/local/etc/mrtg/exec_mrtg.sh > /dev/null 2>&1 > D> >No-no. The question was: what is the scale of time axis? How long does one > D> >peak last? > D> hmm...still difficult to understand your question. well, time axis scale is displayed in hours. each new point is added every 10 minutes. each point value is read immediately > D> by mrtg without any averages or somth. > > So, there is 24 hours between red lines on the image [1]? Then we have 2 peaks > per hour, right? May be this is sendmail (or other MTA) processing its queue? > > [1] ftp://temp:temp@217.9.241.242/hatm0.png your idea seems logical enough....just can't understand why peaks have never been detected on incoming interfaces, before ng_bridge? mrtg doesn't allways reads current values with precission of seconds....for example point 1 - 00:00:15, point 2 - 00:10:37, point 3 - 00:20:24 etc... you might say that incoming traffic on ethernet interfaces is much bigger that incoming traffic bridged to current ngethxxx and it simply swallows peaks, but we've done experiments with only one physical and one logical bridged interfaces - fluctuations have also been noticed. anyway, maybe you have any ideas why after doing "ngctl msg ng_bridge: setconfig '{ ipfw=[1] }', and getconfig showing succesful activation of "ipfw = [ 1 ]", traffic is not passed to ipfw? bsd 5.3... thanks...
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?006101c55d01$8b08fe20$9f90a8c0>