Date: Sat, 29 Jun 2002 23:55:25 -0500 From: Len Conrad <LConrad@Go2France.com> To: freebsd-hackers@freebsd.org Subject: ftp and mail much slower into fbsd 4.4 vs and old BSDi Message-ID: <5.1.0.14.2.20020629235459.031daf28@mail.Go2France.com>
next in thread | raw e-mail | index | archive | help
Sorry, hackers, I posted this twice in -questions and got no response. If the problem is newreno, can somebody say how to up just that piece for 4.4 so as to be as non-disruptive, non-dice-rolling as possible on this otherwise solid machine? Thanks Len ================ FreeBSD 4.4-RELEASE #0 CPU: Pentium III/Pentium III Xeon/Celeron (848.05-MHz 686-class CPU) avail memory = 518156288 (506012K bytes) tx0: <SMC EtherPower II 10/100> port 0x1000-0x10ff mem 0xe8000000-0xe8000fff irq 10 at device 14.0 on pci0 qsphy0: 10baseT, 10baseT-FDX, 100baseTX, 100baseTX-FDX, auto tx0: address 00:e0:29:24:17:80, type SMC9432TX ahc0: <Adaptec 29160 Ultra160 SCSI adapter> port 0x1400-0x14ff mem 0xe8001000-0xe8001fff irq 9 at device 16.0 on pci0 aic7892: Ultra160 Wide Channel A, SCSI Id=7, 32/255 SCBs Mounting root from ufs:/dev/da0s1a da0 at ahc0 bus 0 target 0 lun 0 da0: <SEAGATE ST318437LW 0105> Fixed Direct Access SCSI-3 device da0: 160.000MB/s transfers (80.000MHz, offset 31, 16bit), Tagged Queueing Enabled # sysctl -a | grep tcp tcpcb: 544, 1064, 51, 152, 70748 net.inet.tcp.rfc1323: 1 net.inet.tcp.rfc1644: 0 net.inet.tcp.mssdflt: 512 net.inet.tcp.keepidle: 7200000 net.inet.tcp.keepintvl: 75000 net.inet.tcp.sendspace: 16384 net.inet.tcp.recvspace: 16384 net.inet.tcp.keepinit: 75000 net.inet.tcp.delacktime: 100 net.inet.tcp.v6mssdflt: 1024 net.inet.tcp.log_in_vain: 0 net.inet.tcp.blackhole: 0 net.inet.tcp.delayed_ack: 1 net.inet.tcp.tcp_lq_overflow: 1 net.inet.tcp.path_mtu_discovery: 1 net.inet.tcp.slowstart_flightsize: 1 net.inet.tcp.local_slowstart_flightsize: 65535 net.inet.tcp.newreno: 1 net.inet.tcp.tcbhashsize: 512 net.inet.tcp.do_tcpdrain: 1 net.inet.tcp.pcbcount: 51 net.inet.tcp.icmp_may_rst: 1 net.inet.tcp.strict_rfc1948: 0 net.inet.tcp.isn_reseed_interval: 0 net.inet.tcp.msl: 30000 net.inet.tcp.always_keepalive: 1 # sysctl -a | grep buf kern.ipc.maxsockbuf: 262144 kern.ipc.sockbuf_waste_factor: 8 kern.ipc.mbuf_wait: 32 kern.ipc.nmbufs: 4096 kern.msgbuf: kern.msgbuf_clear: 0 vfs.nfs.bufpackets: 4 vfs.numdirtybuffers: 130 vfs.lodirtybuffers: 499 vfs.hidirtybuffers: 998 vfs.numfreebuffers: 3785 vfs.lofreebuffers: 222 vfs.hifreebuffers: 444 vfs.runningbufspace: 0 vfs.maxbufspace: 64143360 vfs.hibufspace: 63488000 vfs.lobufspace: 63422464 vfs.bufspace: 63422464 vfs.maxmallocbufspace: 3174400 vfs.bufmallocspace: 712704 vfs.getnewbufcalls: 556989 vfs.getnewbufrestarts: 0 vfs.bufdefragcnt: 0 vfs.buffreekvacnt: 0 vfs.bufreusecnt: 3871 vfs.reassignbufcalls: 3716853 vfs.reassignbufloops: 0 vfs.reassignbufsortgood: 474986 vfs.reassignbufsortbad: 1252304 vfs.reassignbufmethod: 1 debug.bpf_bufsize: 4096 debug.bpf_maxbufsize: 524288 machdep.msgbuf: machdep.msgbuf_clear: 0 with a 190 mbyte file: an ftp client pulling the file from fbsd is "blindingly fast" (aka "immeasurably" :)) ) ftp client sending to freebsd is 52 kbytes/sec ftp client sending to another ftp server on the same LAN is 1.6 megabytes/sec sending mail with a large attachment to the fbsd box is 100 kbytes/sec. (postfix is MTA) ftp a 5.5 mbyte from workstation client to fbsd: 109 seconds ftp a 5.5 mbyte from fbsd client to workstation ftp server: 3 seconds the machine runs a apache, qpopper, postfix, ftp, bind9 for a small LAN all on the same segment. dmesg and messages shows no errors. netstat -bi gives Name Mtu Network Address Ipkts Ierrs Ibytes Opkts Oerrs Obytes Coll tx0 1500 <Link#1> 00:e0:29:24:17:80 3171595 552 465267774 4214806 0 720458428 0 yeah, we don't like the Ierrs of 552. change the nic? change the driver? or is this the "newreno" tcp/ip problem? The Ierrs go up by 2 or 3 after each 5.5 mbytes file transfer. The new FBSD box replaces a BSDi on older hardware, with everybody on the LAN now noting that ftp of big files to and mail with big attachments to the new FBSD box are dramatically slower than the BSDi box. Anybody have any idea how to speed up / fix the transfers to the fbsd box? thanks Len www.menandmice.com/DNS-training : DNS Training BIND8NT.MEIway.com : ISC BIND for NT4 & W2K IMGate.MEIway.com : Build free, hi-perf, anti-abuse mail gateways To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-hackers" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?5.1.0.14.2.20020629235459.031daf28>