[K12OSN] Re: Try this utility to test your thin client network

[Terrell Prudé, Jr.]

On Sat, 2005-11-05 at 16:32 -0800, Robert Arkiletian wrote:

> If you don't want to run the program as root then just change the '-f' to a '-A'
> 
> >                 char ping[64]="ping -c100 -f -p aa ";
> 
> Note: If you change it to Adaptive '-A' ping you will get a lot more
> output and the test will take much longer.



Hi Robert,

I also use the flood ping to test my network.  I've found that if I vary
the packet size (the -s parameter), I get a good test.  I'll typically
do 100, 500, and 1490; this last one, with Ethernet frame padding, gives
the standard MTU of 1518 bytes.  The reason for this is that I've found
some poorly-made gear that chokes at an Ethernet frame size beyond 1500
bytes, and also it's helped me to check gear that's starting to fail.
Recently, we had an issue with the Cisco 3750M switch, which was
revealed by doing this.

Also, I'll typically run 1,000 packets per ping session.  I do that
because of a problem similar to one that I had to track down and fix
this last week.  At times, 100-200 pings would get through, but it'd
choke for maybe half a second to five seconds after that.  The problem
turned out to be a malfunctioning wireless access point in a box
somewhere in the school; it was claiming to be the spanning tree root
(yes, I put on rootguard afterwards).

One other thing about both the -A and the -i parameters: my reading
tells me that, if you're not root, the minimum interval you'll be able
to get is one packet every 200ms, is that your experience as well?

--TP
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listman.redhat.com/archives/k12osn/attachments/20051106/3748381f/attachment.htm>