Thanks to:
Joe Shamblim
Daniel Beaudry
Owen Ofiesh
The cause of the problem ended up being the TPE cable (or
so it looks like, so far). I got a couple of interesting
suggestions which I've appended below.
thanks again,
Mariela
ORIGINAL QUESTION:
===========================================================
One of our servers hang last night, and had to be re-setted
(L1-A), as we could not get response from the system.
The messages log file shows lots of the following entries
at, or right before, the time of the failure:
vmunix: data2mbuf: no mbuf for ddp pkt
vmunix: sendETApkt: can not get mbuf..
vmunix: sendETApkt: no mbufs!
vmunix: etput: no mbufs!
vmunix: le0: out of mbufs - packet dropped
The person looking into this issue last night couldn't find
any information re: these errors from Sunsolve.
The server in question is 40 MHz Sun 4/75, with 64MB main mem,
and it is running SunOS 4.1.3
Looking forward to hearing from you,
Mariela
ANSWERS RECEIVED:
===============================================================
================================================================
Is it a SYN flood attack? take a look at:
ftp://info.cert.org/pub/cert_advisories/CA-96.21.tcp_syn_flooding
if you are hooked to the Internet.
Good luck!
Daniel Beaudry
V.P. Technologies Communications MULTIMEDIA & Enter-Net
2100, boul. Rene-Gaultier tel. (514) 652-7189
Varennes (Quebec) fax: (514) 652-6973
J3X 1P1 Daniel.Beaudry@Enter-Net.com
================================================================
What is your MAXUSERS value set to in the config file. If you up this it
will increase the number of mbufs that are alloted. All you need to do is
increase this number, set it to 32 or something, recompile and there you
are. You can also edit the /sys/conf.common/param.c file, and increase
the value there.
Hope that helps,
Joe
Joe Shamblin wjs@cs.duke.edu
=================================================================
Return-Path: <owenf@dlrgatd.com>
Subject: Re: le0: out of mbufs - packet dropped
To: mariela@corp.cirrus.com
Date: Mon, 7 Oct 1996 08:14:05 -0700 (PDT)
Just saw the thread. If you type (as root) on the local console:
ifconfig le0 down
sleep 10
ifconfig le0 up
you should be able to recover the ethernet port with no ill affect to
your system (and with no reboot necessary). You WILL see a lot of error
messages scrolling on the console complaining about "bad send and reply
packets" (it's been a while, so I don't remember the exact format of
message off hand - Sorry.).
The problem will re-occur until you figure out what's at the source of
it all. If you need help with that, I have some ideas. Drop me a line.
BTW: This is a good argument for not having a root path that accesses
anything via any nfs/network path...
Good luck...
================================================================
This archive was generated by hypermail 2.1.2 : Fri Sep 28 2001 - 23:11:11 CDT