From owner-freebsd-net Tue May 5 07:35:19 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id HAA24873 for freebsd-net-outgoing; Tue, 5 May 1998 07:35:19 -0700 (PDT) (envelope-from owner-freebsd-net@FreeBSD.ORG) Received: from jingoro.prevmed.sunysb.edu (jingoro.prevmed.sunysb.edu [129.49.123.10]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id HAA24855 for ; Tue, 5 May 1998 07:35:04 -0700 (PDT) (envelope-from cfanning@jingoro.prevmed.sunysb.edu) Received: (from cfanning@localhost) by jingoro.prevmed.sunysb.edu (8.8.8/8.8.5) id KAA01507 for freebsd-net@freebsd.org; Tue, 5 May 1998 10:26:08 GMT From: Chris Fanning Message-Id: <199805051026.KAA01507@jingoro.prevmed.sunysb.edu> Subject: sudden MBUF problem To: freebsd-net@FreeBSD.ORG Date: Tue, 5 May 1998 10:26:08 +0000 (GMT) X-Mailer: ELM [version 2.4ME+ PL31H (25)] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: owner-freebsd-net@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org All was fine and good with FreeBSD's networking (for years!) until this last weekend. I first noticed the problem dialed in - frequent disconnection, packets would get delayed for a crazy amount of time, etc. In fact, it's happening right now, but only a lot less often. It's not easy diagnosing a problem remotely when you get kicked out every minute! I found this in my syslog: May 5 09:54:04 jingoro /kernel: Out of mbuf clusters - increase maxusers! So I searched for "mbuf clusters" and found a kernel setting to increase them. All was fine with the standard settings until this weekend, so I was a bit puzzled. Nothing changed on the machine and it gets very light loading. The output of my "netstat -m" is: 4704 mbufs in use: 4434 mbufs allocated to data 4 mbufs allocated to packet headers 9 mbufs allocated to protocol control blocks 253 mbufs allocated to fragment reassembly queue headers 4 mbufs allocated to socket names and addresses 2248/2262 mbuf clusters in use 5112 Kbytes allocated to network (99% in use) 0 requests for memory denied 15 requests for memory delayed 15 calls to protocol drain routines And my uptime right now is less than 20 minutes. I'm guessing something's going on where I'm getting a crazy amount of mbufs being allocated to data for no good reason. I was running 2.2-STABLE and upgraded to 2.2.6-STABLE with the same results. Could I be the victim of some degenerative TCP/IP attack? I ran Ethload on my network and found no packets being specifically destined for my MAC... After a protocol drain, the mbuf count in use climbs a few hundred per second until it gets to the top where it stays for a while. Anyone have any ideas on this? Or what I can do to determine what's causing the problem? Thanks, Chris To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-net" in the body of the message