doug@csd4.milw.wisc.edu (Doug Tiarks) (05/17/89)
Awhile back douglis@MINT.BERKELEY.EDU (Fred Douglis) reported a problem with sendmail: >>We're experiencing a problem sending mail to an internet host >>(decwrl.dec.com) -- the symptom is that the mailer hangs for a long >>time and then complains "reply: read error", apparently in the phase >>just after it sends the body of the message. I heard that about the >>same time we started having this problem on Sprite, a small number of >>Suns running SunOS 3.2 had the same problem, and I was told that >>upgrading the kernel to 3.4 fixed the problem. This is why I believe >>the problem might have something to do with the TCP connection (since >>just about everything at user level is the same, but the kernel is >>new). We are apparently seeing this problem from the other side. Occasionally sendmail will go to sleep (waiting for input ?) during a SMTP session with another host which is trying to deliver a message to us. This occurs at different stages (HELO, MAIL FROM, DATA) of the SMTP session. Sendmail eventually times out, but the host on the other end apparently makes multiple tries during the timeout period with the same result. We're running v5.61 of sendmail with some fixes from comp.bugs.4bsd under 4.3bsd-tahoe on a tahoe machine. Anyone experiencing similar problems? Anyone know of a fix? Doug Tiarks UW-Milwaukee Computing Services Div. Internet: doug@csd4.milw.wisc.edu UUCP: {uwvax, uwmacc}!uwmcsd1!doug -- Doug Tiarks UW-Milwaukee Computing Services Div. Internet: doug@csd4.milw.wisc.edu UUCP: {uwvax, uwmacc}!uwmcsd1!doug
prue@VENERA.ISI.EDU (05/20/89)
Doug, Jim Guyton at Rand described this same problem to me. He and I did some poking around and found that somewhere between here (Southern Calif.) and decwrl.dec.com, someone is clobbering packets larger than somewhere in the 1000-->1100 byte range. We can ping all day with packets less than 1000 and get good reliablility but 1100 or more fails consistently. Walt Prue
"Juan_Navarro.XOSMAR"@XEROX.COM (05/22/89)
Doug- When I was connecting our (XEROX's) TCP/IP gateway to Sun OS3.x (.2 or .5) we were having the same problem. We fixed it by adding a E=\r\n at the end of the Mtcp or Mether line in the sendmail.cf file. I forget where this is documented, but it works. You then need to kill and restart your sendmail daemon. Example: Mtcp, P=[IPC], F=mDFMueXL, S=14, R=14, A=IPC $h, E=\r\n or Mether, P=[IPC], F=mDFMueXL, S=11, R=21, A=IPC $h, E=\r\n Good Luck. //juan
dsmith@oregon.uoregon.edu (Dale Smith) (05/25/89)
While we are on the subject of mail, is anyone else having terrible trouble delivering mail to the UK? I have have had consistent problems for some time now (6 weeks?). I ignored it for awhile, but this discussion got me to poking about and it looks like someone between me and nsfnet-relay.ac.uk is dropping packets longer than 500 bytes or so. Note that the packets are not getting fragmented, they are getting dropped. Now, I could crank down the MTU on my mail host, but I also use it for all kinds of other things and that would kill me for performance. Any advice? Pointers? Anyone who has no trouble willing to act as a mail forwarder for me for .uk mail? At the end of this message is a traceroute between myself and nsfnet-relay.ac.uk. (128.86.8.6). Thanks, Dale Smith Asst Dir of Network Services Internet: dsmith@oregon.uoregon.edu University of Oregon BITNET: dsmith@oregon.bitnet Computing Center UUCP: ...hp-pcd!uoregon!dsmith Eugene, OR 97403-1212 Voice: (503)686-4394 ---------------------------------------------------------------------- hogg# traceroute 128.86.8.6 traceroute to 128.86.8.6 (128.86.8.6), 30 hops max, 38 byte packets 1 uo-gw.uoregon.edu (128.223.20.1) 20 ms 0 ms 20 ms 2 192.33.18.10 (192.33.18.10) 500 ms 200 ms 180 ms 3 192.31.215.10 (192.31.215.10) 560 ms 560 ms 380 ms 4 129.101.101.10 (129.101.101.10) 1260 ms 580 ms 980 ms 5 192.31.216.10 (192.31.216.10) 1000 ms 640 ms 660 ms 6 192.35.180.3 (192.35.180.3) 700 ms 2500 ms 2020 ms 7 129.140.70.14 (129.140.70.14) 400 ms 400 ms 460 ms 8 129.140.71.6 (129.140.71.6) 440 ms 1200 ms 460 ms 9 129.140.81.7 (129.140.81.7) 500 ms 740 ms 480 ms 10 129.140.72.17 (129.140.72.17) 560 ms 1100 ms 740 ms 11 * * * 12 ford-gateway.csc.org (128.121.54.73) 600 ms 680 ms 580 ms 13 128.121.54.78 (128.121.54.78) 580 ms 740 ms 540 ms 14 * * * 15 * * * 16 * * * 17 * * * 18 * * * 19 * * * 20 * * * 21 * * * 22 * * * 23 * * * 24 * * * 25 * * * 26 * * * 27 * NSFNET-RELAY.AC.UK (128.86.8.6) 1600 ms 1660 ms