Date: Tue, 12 Jul 2005 16:49:25 +0300 From: Giorgos Keramidas <keramida@freebsd.org> To: Roman Kurakin <rik@cronyx.ru> Cc: freebsd-net@freebsd.org Subject: Re: ping delay, initial request Message-ID: <20050712134925.GB1061@beatrix.daedalusnetworks.priv> In-Reply-To: <42D3B9A8.6000803@cronyx.ru> References: <42D3B9A8.6000803@cronyx.ru>
next in thread | previous in thread | raw e-mail | index | archive | help
On 2005-07-12 16:38, Roman Kurakin <rik@cronyx.ru> wrote: > Hi, > > I am starting to investigate some net problem and I wonder if this > problem seen/known. > The problem was observed with ce(4) (Cronyx Tau32-PCI/Lite, > it is not in the tree yet, but similar to cp(4)/ctau(4) devices) and > sppp(4). > If you run usual ping you will see normal delay which is much less than > 1sec. > But if you run flood ping, stop it, and run normal ping again you'll see > delay > about one sec. > > ping x.x.x.x > delay << 1 sec > ping -f x.x.x.x > average delay << 1 sec > ping x.x.x.x > delay ~ 1 sec > > This was seen on FreeBSD 4.11 stable. > > Any ideas? Does any body observed such behaviour in other environment? Is it possible that flood ping hits an icmp rate-limiting watermark and then every subsequent icmp packet gets penalized with a delay until a fair amount of time passes?
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20050712134925.GB1061>