Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!watmath!clyde!burl!mgnetp!ihnp4!mhuxl!houxm!houxz!vax135!cornell!uw-beaver!tektronix!hplabs!sri-unix!mike@BRL-TGR.ARPA
From: mike@BRL-TGR.ARPA
Newsgroups: net.unix-wizards
Subject: Re:  Talk Command
Message-ID: <2061@sri-arpa.UUCP>
Date: Mon, 16-Jul-84 15:58:50 EDT
Article-I.D.: sri-arpa.2061
Posted: Mon Jul 16 15:58:50 1984
Date-Received: Fri, 20-Jul-84 04:03:35 EDT
Lines: 40

From:      Mike Muuss <mike@BRL-TGR.ARPA>

[4.2 BSD network trivia]
"connect:  out of buffer space" generally does NOT mean that
the kernel is out of MBUFS (easily checked with "netstat -m"),
but instead usually means that there is a backlog in the queue
destined for your output device and that packets are being
dropped.

The most common case of this is with 1822 IMP interfaces connected
via ECUs, under 2 conditions:

1)  ECUs have not been set to "15 retry, no relay flapping" mode
(all retry switches up, rightmost switch in that row DOWN).

2)  ECUs are operating on a line with LONG noise bursts.  (We have 1
such line that goes through a swamp...).  What happens here is that
RFNM messages from the IMP are lost, and it confuses the 1822
link optimization code.  (Check this with "netstat -h", look for
persistant RFNM counts > 2 ).  I have patches for /sys/netinet/if_imp.c
that can be activated with "options NORFNM" in the config file,
drop me a line if you need it.

Best,
 -Mike Muuss

(301)-278-6678
  AV  283-6678
  FTS 939-6678

ArpaNet:  Mike @ BRL
UUCP:     ...!{decvax,cbosgd}!brl-bmd!mike
Postal:
  Mike Muuss
  Leader, Advanced Computer Systems Team
  Computer Techniques and Analysis Branch
  Systems Engineering and Concepts Analysis Division
  U.S. Army Ballistic Research Laboratory
  Attn: DRXBR-SECAD (Muuss)
  APG, MD  21005