From owner-freebsd-questions Sun Nov 16 08:50:29 1997 Return-Path: Received: (from root@localhost) by hub.freebsd.org (8.8.7/8.8.7) id IAA29785 for questions-outgoing; Sun, 16 Nov 1997 08:50:29 -0800 (PST) (envelope-from owner-freebsd-questions) Received: from woman.kitel.co.kr (woman.kitel.co.kr [210.116.210.106]) by hub.freebsd.org (8.8.7/8.8.7) with ESMTP id IAA29775 for ; Sun, 16 Nov 1997 08:50:21 -0800 (PST) (envelope-from swjeong@woman.kitel.co.kr) Received: from woman.kitel.co.kr (localhost.kitel.co.kr [127.0.0.1]) by woman.kitel.co.kr (8.8.7/8.8.7) with SMTP id BAA00438; Mon, 17 Nov 1997 01:50:00 +0900 (KST) (envelope-from swjeong@woman.kitel.co.kr) Message-ID: <346F2438.41C67EA6@woman.kitel.co.kr> Date: Mon, 17 Nov 1997 01:50:00 +0900 From: Jeong Seong Won Organization: kitel X-Mailer: Mozilla 3.01 (X11; I; FreeBSD 2.2.5-RELEASE i386) MIME-Version: 1.0 To: dg@root.com CC: freebsd-questions@FreeBSD.ORG Subject: Re: Network connection often stops. Why ? References: <199711160800.AAA12641@implode.root.com> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Sender: owner-freebsd-questions@FreeBSD.ORG X-Loop: FreeBSD.org Precedence: bulk > > >In my company. About twenty of 686 FreeBSD 2.2.1 Machine is connected > > via FastEthernet.(Intel EtherExpress) > > Hub is SMC tigerstack.(not a switching) > > Problem occurs in the same way when I used SMC 9332BDT. > > Network options MAXMEM=262144 > > Concurrent User is aboue 230 per machine. > > > > Problem is that suddenly telnet session stop. > > ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ > > When this happens, all key stokes is not effective, and at last > > conection is closed. > > > > vnode, file table, mbuf, somaxconn is sufficient. > > > > kernel configuration is > > > >.... > > > >options MAXMEM=262144 > >maxusers 250 > >options NMBCLUSTERS=8192 > >options > I just replied to your message in the NetBSD list as well. > I think the problem is that you are running out of mbuf clusters > (even though you have the setting really high already). > Can you monitor that with "netstat -m" and > make sure that the second figure in the ratio of mbuf clusters > in use is never more than 6000? I make sure mbuf never more than 6000. The problem often occurs when no one except root on console logined. (when 'cd /;ls -lR') Below is system status (netstat -s, netstat -m) when it's load is relatively middle.(156 user) Thank you. ---------------------------------------------------------------- $ netstat -m 1521 mbufs in use: 99 mbufs allocated to data 658 mbufs allocated to packet headers 762 mbufs allocated to protocol control blocks 2 mbufs allocated to socket names and addresses 40/1112 mbuf clusters in use 2414 Kbytes allocated to network (11% in use) 0 requests for memory denied 0 requests for memory delayed 0 calls to protocol drain routines Here is system status. (netstat -m, netstat -s) $ netstat -s (uptime 2 days, 17h) ip: 70806126 total packets received 0 bad header checksums 0 with size smaller than minimum 0 with data size < data length 0 with header length < data size 0 with data length < header length 0 with bad options 0 with incorrect version number 0 fragments received 0 fragments dropped (dup or out of space) 0 fragments dropped after timeout 0 packets reassembled ok 68037217 packets for this host 2768909 packets for unknown/unsupported protocol 0 packets forwarded 0 packets not forwardable 0 redirects sent 62560291 packets sent from this host 0 packets sent with fabricated ip header 0 output packets dropped due to no bufs, etc. 0 output packets discarded due to no route 0 output datagrams fragmented 0 fragments created 0 datagrams that can't be fragmented icmp: 2769059 calls to icmp_error 0 errors not generated 'cuz old message was icmp Output histogram: echo reply: 5 destination unreachable: 2769059 0 messages with bad code fields 0 messages < minimum length 0 bad checksums 0 messages with bad length Input histogram: echo reply: 14 destination unreachable: 2771433 echo: 5 5 message responses generated igmp: 0 messages received 0 messages received with too few bytes 0 messages received with bad checksum 0 membership queries received 0 membership queries received with invalid field(s) 0 membership reports received 0 membership reports received with invalid field(s) 0 membership reports received for groups to which we belong 0 membership reports sent tcp: 56960230 packets sent 39981139 data packets (-1116894515 bytes) 43014 data packets (20673340 bytes) retransmitted 0 resends initiated by MTU discovery 9784510 ack-only packets (6816381 delayed) 6 URG only packets 7847 window probe packets 4767211 window update packets 2378547 control packets 65223827 packets received 34333052 acks (for -1115532635 bytes) 897408 duplicate acks 0 acks for unsent data 41680061 packets (-323558872 bytes) received in-sequence 49030 completely duplicate packets (33264557 bytes) 0 old duplicate packets 414 packets with some dup. data (168205 bytes duped) 1232106 out-of-order packets (507712999 bytes) 2811 packets (4128 bytes) of data after window 1387 window probes 414935 window update packets 7351 packets received after close 16 discarded for bad checksums 0 discarded for bad header offset fields 0 discarded because packet too short 1165552 connection requests 68821 connection accepts 323 bad connection attempts 0 listen queue overflows 1218177 connections established (including accepts) 1233896 connections closed (including 3712 drops) 61972 connections updated cached RTT on close 61972 connections updated cached RTT variance on close 43499 connections updated cached ssthresh on close 12719 embryonic connections dropped 30767651 segments updated rtt (of 30729917 attempts) 40866 retransmit timeouts 675 connections dropped by rexmit timeout 7963 persist timeouts 3 connections dropped by persist timeout 3116 keepalive timeouts 1332 keepalive probes sent 163 connections dropped by keepalive 14195968 correct ACK header predictions 28492503 correct data packet header predictions udp: 2810847 datagrams received 0 with incomplete header 0 with bad data length field 0 with bad checksum 2769059 dropped due to no socket 5124 broadcast/multicast datagrams dropped due to no socket 0 dropped due to full socket buffers 0 not for hashed pcb 36664 delivered 2805863 datagrams output $ pstat -T 1932/8040 files 3763 vnodes 16M/139M swap space