Dropping packets??

[ME]

This makes no sense when I read my own post. Revisited:

On Fri, 1 Feb 2002, ME wrote:

> Date: Fri, 1 Feb 2002 18:03:15 -0800 (PST)
> From: ME <dugan at passwall.com>
> Reply-To: talk at nblug.org
> To: "'talk at nblug.org'" <talk at nblug.org>
> Subject: RE: Dropping packets??
> 
> On Fri, 1 Feb 2002, Christopher White wrote:
> > > I tried pinging www.yahoo.com, www.whitehouse.gov, 
> > > www.nomaddesign.com,
> > > thunder.sonic.net, newbolt.sonic.net, etc..  All seem to give 
> > > me some packet
> > > loss..  Except for your IP..  0% packet loss.  It seems 
> > > though, that the
> > > packet loss is lower now than before now that I've been 
> > > messing around with
> > > it..  Weird..
> 
> I have not beeing keeping track of this thread.
             been
> 
> You realize that when you ping someone, if you interrupt the ping process
> after you sent a ping, but before you received the response, it will say
> you have packet loss, correct?
> 
> If you uset large packet sizes for ping then that can lead to return
> packets return trip and timestamp exceeding window size used in the
ping process and have it count the sent packet as missed when internally
if it is later received out of sequence, it may sho less packet loss in 
the summary than the interactive display suggests.

> 
> If the above, try your ping as before, but instead use the following
> flags:
> 
> $ ping -s 56 -c 5 host.to.ping
> 
> This should allow you to ping without interruptions between packets and 
only send/receive 5 ping packets and then stop on its own (do not press
control-c for this or interrupt it.)

> (Side note: If you ping, and send 2 pings, but only receive 1 since you
> interrupt the ping process after you got the response for one but the send
> for the other, then you may see 50% packet loss, when in reality that is
> just a matter of un received return caused by interruption of the program.
> 
> Could you send the original post direct to me (forward it) as I may have
> deleted it thinkin it was part of another thread.