ping: sendto: No buffer space available

[jaime at snowmoon.com]

On Tue, 17 Jun 2003, lbland wrote:
> I don't know, but it may be a router loop problem in the ISP router
> tables. Those tables can change dynamically and can cause intermittent
> issues like you explained.

	I had three pings going at the same time.  One to the ISP's DNS
resolver, one to the far end of the T-1, and one to the ethernet interface
on the router at my site.  The router and firewall are on opposite ends of
the same cable.

	When the pings to the DNS resolver gave the "No buffer space"
message, so did the other two pings.  This means that the break down is
not any further up stream than the router.  I'm now running pings to a
host on the same LAN as the firewall.  The next time that the "No buffer
space" message appears on the pings to the DNS resolver, I'll check the
pings to the internal host.  If they have the same problem, then I'm
experiencing an OS level issue of some kind.

	OK, it happened while I was typing this.  :)  Results:  internal
host remained ping-able while the other three pings were all giving "No
buffer space" messages.

	This is starting to sound like some kind of packet over-load on
the "public" side of my FreeBSD/ipfw based firewall.  Does anyone have any
advice on how to confirm this?

bash-2.05b$ uname -a
FreeBSD cerberus.cairodurham.org. 4.7-STABLE FreeBSD 4.7-STABLE #0: Sat
Oct 12 12:54:03 EDT 2002     jkikpole at cerberus.cairodurham.org.:/usr/obj/usr/src/sys/CERBERUS  i386

							Thanks in advance,
							Jaime