qmail Digest 14 May 1999 10:00:00 -0000 Issue 640
Topics (messages 25526 through 25596):
Embedded linefeed epidemic
25526 by: Chris Johnson <[EMAIL PROTECTED]>
25572 by: Scott Schwartz <[EMAIL PROTECTED]>
Qmail crashed....
25527 by: �����ӿ��� <[EMAIL PROTECTED]>
25529 by: Russell Nelson <[EMAIL PROTECTED]>
25530 by: Dave Sill <[EMAIL PROTECTED]>
25532 by: Dave Sill <[EMAIL PROTECTED]>
25534 by: Yessure <[EMAIL PROTECTED]>
setting relay clients
25528 by: Russell Nelson <[EMAIL PROTECTED]>
script to tail latest logfile in a directory?
25531 by: "Robin Bowes" <[EMAIL PROTECTED]>
25535 by: Dave Sill <[EMAIL PROTECTED]>
25541 by: "Robin Bowes" <[EMAIL PROTECTED]>
25545 by: "Scott D. Yelich" <[EMAIL PROTECTED]>
25546 by: Dave Sill <[EMAIL PROTECTED]>
25547 by: "Robin Bowes" <[EMAIL PROTECTED]>
25548 by: "Scott D. Yelich" <[EMAIL PROTECTED]>
25549 by: "Scott D. Yelich" <[EMAIL PROTECTED]>
25551 by: Dave Sill <[EMAIL PROTECTED]>
25552 by: "Robin Bowes" <[EMAIL PROTECTED]>
25561 by: Jeff Hayward <[EMAIL PROTECTED]>
25562 by: "Scott D. Yelich" <[EMAIL PROTECTED]>
25563 by: "Adam D. McKenna" <[EMAIL PROTECTED]>
25566 by: "Adam D. McKenna" <[EMAIL PROTECTED]>
QMAIL definitely violates PIPELINING specification ...
25533 by: Russell Nelson <[EMAIL PROTECTED]>
25556 by: "Fred Lindberg" <[EMAIL PROTECTED]>
Qmail to procmail and back again?
25536 by: Stig Sandbeck Mathisen <[EMAIL PROTECTED]>
25583 by: [EMAIL PROTECTED]
forwarding question
25537 by: Dave Sill <[EMAIL PROTECTED]>
Mail Queue Just Keeps Growing....
25538 by: "Christopher Porreca" <[EMAIL PROTECTED]>
25540 by: Dave Sill <[EMAIL PROTECTED]>
25543 by: Gerry Boudreaux <[EMAIL PROTECTED]>
25553 by: Jere Cassidy <[EMAIL PROTECTED]>
25557 by: "Christopher Porreca" <[EMAIL PROTECTED]>
25559 by: Kevin Sawyer <[EMAIL PROTECTED]>
25560 by: Dave Sill <[EMAIL PROTECTED]>
25564 by: xs <[EMAIL PROTECTED]>
qmail config questions
25539 by: Dave Sill <[EMAIL PROTECTED]>
25542 by: "Scott D. Yelich" <[EMAIL PROTECTED]>
Could someone help, please?
25544 by: Dave Sill <[EMAIL PROTECTED]>
help: tcpserver dies.
25550 by: Jhirley Fonte <[EMAIL PROTECTED]>
25554 by: Dave Sill <[EMAIL PROTECTED]>
Fundamental flaws in List-Unsubscribe
25555 by: "Fred Lindberg" <[EMAIL PROTECTED]>
25565 by: Mate Wierdl <[EMAIL PROTECTED]>
25585 by: "D. J. Bernstein" <[EMAIL PROTECTED]>
25586 by: "Fred Lindberg" <[EMAIL PROTECTED]>
blocking mail send/receive from a domain?
25558 by: Bill Parker <[EMAIL PROTECTED]>
25571 by: Russell Nelson <[EMAIL PROTECTED]>
25573 by: Tillman <[EMAIL PROTECTED]>
25574 by: Russell Nelson <[EMAIL PROTECTED]>
sendmail/qmail doesn't work
25567 by: Luca Pescatore <[EMAIL PROTECTED]>
taildir won't compile...
25568 by: Marlon Anthony Abao <[EMAIL PROTECTED]>
25575 by: Jeff Hayward <[EMAIL PROTECTED]>
25580 by: "Adam D. McKenna" <[EMAIL PROTECTED]>
25581 by: Justin Bell <[EMAIL PROTECTED]>
25582 by: "Scott D. Yelich" <[EMAIL PROTECTED]>
25584 by: Scott Schwartz <[EMAIL PROTECTED]>
25587 by: Marlon Anthony Abao <[EMAIL PROTECTED]>
25591 by: Troy Morrison <[EMAIL PROTECTED]>
25592 by: "Scott D. Yelich" <[EMAIL PROTECTED]>
25594 by: "Robin Bowes" <[EMAIL PROTECTED]>
Qmail + automail setup
25569 by: Logics <[EMAIL PROTECTED]>
Different logs for different domains
25570 by: Stathakopoulos Giorgos <[EMAIL PROTECTED]>
25577 by: Anand Buddhdev <[EMAIL PROTECTED]>
Mail Queue Fixed!
25576 by: "Christopher Porreca" <[EMAIL PROTECTED]>
interesting bounce
25578 by: Peter van Dijk <[EMAIL PROTECTED]>
[[EMAIL PROTECTED]: (fwd) Re: Zmailer and Qmail are not fully compatible ...]
25579 by: Peter van Dijk <[EMAIL PROTECTED]>
Email sending problem
25588 by: Michael Mansour <[EMAIL PROTECTED]>
25589 by: [EMAIL PROTECTED]
25590 by: Chris Johnson <[EMAIL PROTECTED]>
25593 by: Scott Schwartz <[EMAIL PROTECTED]>
Qmail with Debian with Autoturn and rcpthosts
25595 by: Gavin Lewandowski <[EMAIL PROTECTED]>
25596 by: Anand Buddhdev <[EMAIL PROTECTED]>
Administrivia:
To subscribe to the digest, e-mail:
[EMAIL PROTECTED]
To unsubscribe from the digest, e-mail:
[EMAIL PROTECTED]
To bug my human owner, e-mail:
[EMAIL PROTECTED]
To post to the list, e-mail:
[EMAIL PROTECTED]
----------------------------------------------------------------------
On Thu, May 13, 1999 at 05:03:01AM -0000, D. J. Bernstein wrote: > Chris Johnson writes: > > I'm seeing more and more of qmail-smtpd exiting with status 256, which > > usually indicates that the remote host is trying to send mail with > > bare linefeeds. > > Maybe, maybe not. I didn't realize anybody was trying to extract useful > information from the exit code of qmail-smtpd! There's one increasingly > common situation, having nothing to do with bare linefeeds, in which > qmail-smtpd will exit nonzero; this isn't something to worry about. I realize that a non-zero exit code from qmail-smtpd doesn't necessarily mean that the remote host is sending bare linefeeds, but when I see "status 256" repeatedly from a particular host and then take steps to pipe messages from that host through fixcr before qmail-smtpd, mail from that host is usually (but not always) received successfully. I'd guess that about three quarters of the time this works. What's the "increasingly common situation, having nothing to do with bare linefeeds, in which qmail-smtpd will exit nonzero"? Chris
"D. J. Bernstein" <[EMAIL PROTECTED]> writes: | If you want to watch for particular error messages from qmail-smtpd, run | it under recordio. Make sure you filter the log appropriately---recordio | produces a lot of output. Better yet, patch qmail-smtpd to do sensible logging. Then you get exactly the output you want and need.
Hi, Do you have apache 1.3.6? There is a benchmark program 'ab',and you can run it:(src/support/ab) ./ab -c 800 yourhost:25/abc.htm then qmail will crash....(btw,yourhost:80 SOMETIME can cause apache crash ,but qmail ALWAYS... :( so I dont't think it is because of os problem) I dont't know the reason,but maybe too many connections will cause qmail crash? and how can I resolve this problem? Thanks. Yessure
Could you describe "crash" in more detail? Are you using tcpserver or inetd? �����ӿ��� writes: > Hi, > > Do you have apache 1.3.6? There is a benchmark program 'ab',and you can run > it:(src/support/ab) > ./ab -c 800 yourhost:25/abc.htm > > then qmail will crash....(btw,yourhost:80 SOMETIME can cause apache crash ,but >qmail > ALWAYS... :( so I dont't think it is because of os problem) > > I dont't know the reason,but maybe too many connections will cause qmail crash? and > how can I resolve this problem? > > Thanks. > > > > Yessure > >
�����ӿ��� <[EMAIL PROTECTED]> wrote: > > Do you have apache 1.3.6? There is a benchmark program 'ab',and you can run >it:(src/support/ab) > ./ab -c 800 yourhost:25/abc.htm > > then qmail will crash....(btw,yourhost:80 SOMETIME can cause apache crash ,but qmail >ALWAYS... :( so I dont't think it is because of os problem) It didn't crash any of the three qmails I tested here. -Dave
Russell Nelson <[EMAIL PROTECTED]> wrote: > >Could you describe "crash" in more detail? >Are you using tcpserver or inetd? Dollars to doughnuts he's using inetd and "crash" means connections to port 25 are refused. Dan, is it time to declare inetd unsupported, yet? -Dave
Oh,you are right. :) I just found the log in freebsd's messages but not maillog,and it show: May 4 12:33:40 webmail1 inetd[144]: smtp/tcp server failing (looping), service terminated Now I know how to do, thank you very much. >Russell Nelson <[EMAIL PROTECTED]> wrote: >> >>Could you describe "crash" in more detail? >>Are you using tcpserver or inetd? > >Dollars to doughnuts he's using inetd and "crash" means connections to >port 25 are refused. > >Dan, is it time to declare inetd unsupported, yet? > >-Dave Yessure [EMAIL PROTECTED]
Oden Eriksson writes: > Yes the Russ (or who ever) is wrong about this at the > www.qmail.org site, at least RH5.1-RH6 has tcp wrappers built "the > right way". I think Russ Nelson should update the info about this ? And Redhat is "most Linux distributions"? -- -russ nelson <[EMAIL PROTECTED]> http://crynwr.com/~nelson Crynwr supports Open Source(tm) Software| PGPok | There is good evidence 521 Pleasant Valley Rd. | +1 315 268 1925 voice | that freedom is the Potsdam, NY 13676-3213 | +1 315 268 9201 FAX | cause of world peace.
Hi, I often find that I need to tail the most recent logfile in my qmail log directory ie /var/log/qmail. At the moment, I just look for the most recent file and use (for example): $ tail -f /var/log/qmail/\@00000926439774 | tailocal However, those file names are a bit of a 'mare to type accurately. bash filename completion helps a little but is not really a satisfactory solution. Before I do a bit of coding, has anyone written a script to identify the most recent log file and tail it, preferably switching files when the log file turns over? Ta, R. -- Robin Bowes - System Development Manager - Room 405A E.O.C., Overseas House, Quay St., Manchester, M3 3HN, UK. Tel: +44 161 838 8321 Fax: +44 161 835 1657
"Robin Bowes" <[EMAIL PROTECTED]> wrote: > >Before I do a bit of coding, has anyone written a script to identify the >most recent log file and tail it, preferably switching files when the >log file turns over? Jeff Hayward's taildir does what you want. It's small, and I don't have a URL, so I've attached a copy. -Dave
taildir.c
Dave Sill wrote: > > "Robin Bowes" <[EMAIL PROTECTED]> wrote: > > > >Before I do a bit of coding, has anyone written a script to identify the > >most recent log file and tail it, preferably switching files when the > >log file turns over? > > Jeff Hayward's taildir does what you want. It's small, and I don't > have a URL, so I've attached a copy. Bingo! I was sure I wouldn't be the only person needing to do this. Thanks Dave (and Jeff!) BTW, being a bit of a clueless neophyte when it comes to building C programs, is this an appropriate way to build the file: $ egcc taildir.c -o taildir The resulting binary seems to work so I would conclude that it is! R. -- Robin Bowes - System Development Manager - Room 405A E.O.C., Overseas House, Quay St., Manchester, M3 3HN, UK. Tel: +44 161 838 8321 Fax: +44 161 835 1657
-----BEGIN PGP SIGNED MESSAGE----- On 13 May 1999, Robin Bowes wrote: > > >Before I do a bit of coding, has anyone written a script to identify the > > >most recent log file and tail it, preferably switching files when the > > >log file turns over? > > Jeff Hayward's taildir does what you want. It's small, and I don't > > have a URL, so I've attached a copy. > Bingo! I was sure I wouldn't be the only person needing to do this. What's wrong with something like: tail -f `ls -rt | tail -1` The only thing I wish tail would do is multi-tail... mtail... tail -multi, tail file file file... etc. Scott -----BEGIN PGP SIGNATURE----- Version: 2.6.2 iQCVAwUBNzrQgB4PLs9vCOqdAQGoxgP+N6R9DwznYrQ+ht2RD9WpMxiqkXUqf8ke DD/J+J/Qbgs3tLK7ncI5ttNwedseJn46ubep30PxsGdYnW20yTrtJJVVR9/iJUEo xuSLaa/idnDeC3xElHjpxAomgS8GJGpP4ZaxL0hGB+hIDB9xQjwLO/ubE8lk0Jul BCxIGrIyXr8= =VN2G -----END PGP SIGNATURE-----
"Scott D. Yelich" <[EMAIL PROTECTED]> wrote: > >What's wrong with something like: tail -f `ls -rt | tail -1` Doesn't notice when a new file is created, which cyclog is want to do. -Dave
"Scott D. Yelich" wrote: > > What's wrong with something like: tail -f `ls -rt | tail -1` Erm, nothing. Except that this won't switch files when the log rolls over. `ls -rt | tail -l' will only be executed once at start up. Besides, "taildir" is less typing than "tail -f `ls -rt | tail -1`" :o) > The only thing I wish tail would do is multi-tail... mtail... > tail -multi, tail file file file... etc. You mean monitor several files at once? R. -- Robin Bowes - System Development Manager - Room 405A E.O.C., Overseas House, Quay St., Manchester, M3 3HN, UK. Tel: +44 161 838 8321 Fax: +44 161 835 1657
-----BEGIN PGP SIGNED MESSAGE----- On Thu, 13 May 1999, Dave Sill wrote: > "Scott D. Yelich" <[EMAIL PROTECTED]> wrote: > >What's wrong with something like: tail -f `ls -rt | tail -1` > Doesn't notice when a new file is created, which cyclog is want to do. > -Dave Um... so is the trigger a new file has been created or no more input is being put into the file? Each one of those begs a question... the first is, do you want to "tail" or "cat" each and every new file that is generated and the other is "just when do you know that no more input has been received?" It wouldn't be all that difficult to write a command that came close to doing those although they seem arbitrarily complicated. Scott -----BEGIN PGP SIGNATURE----- Version: 2.6.2 iQCVAwUBNzrSrR4PLs9vCOqdAQEvIgP9HJnqoa7vui2MQj0GCfkQQwJ0fHCLoNl5 ANDd/vXwfuutvXwWbWNSKxWp7JQIW1ZFn2oZDGAnEAyF/O9BsFO0VZvHOFS0snom 2YsGI32UpAhEW48YV/Brh7tznkNvSjAoNYm/dxbUUVmoeBEhmHr177t4q09pz1dW k9n4zFUXp4A= =dmAy -----END PGP SIGNATURE-----
-----BEGIN PGP SIGNED MESSAGE----- On 13 May 1999, Robin Bowes wrote: > Erm, nothing. Except that this won't switch files when the log rolls > over. `ls -rt | tail -l' will only be executed once at start up. Right, sorry all... I was just trying to help. I didn't notice the need to type in only one command once that would change files in the original question. > Besides, "taildir" is less typing than "tail -f `ls -rt | tail -1`" Yup... but now where do I get taildir? Is this more DBWare? Etc. I'm not sure I know the erason why its necessary to "switch" files or what the trigger is -- but I'm sure there is a "standard" unix way to accomplish this. > > The only thing I wish tail would do is multi-tail... mtail... > > tail -multi, tail file file file... etc. > You mean monitor several files at once? Yes. mtail -f /var/adm/messages /var/log/syslog /bing/bang/boom /yadda/yadda/yadda /etc/etc/etc. Scott ps: I'll shut up now... :-/ -----BEGIN PGP SIGNATURE----- Version: 2.6.2 iQCVAwUBNzrTvx4PLs9vCOqdAQGVUAP/Uc9p+Slakxxz+/QiJpETUoRDkmyU8VMs Pnmt4hUAOkBwq5Nb6TIQyKavW4Vm+2v20WbaI28ydWzNceiv1Miavg6M498Dgie8 4v75X0W04hc+Ho9nVEOUNreW9PGhn4QqtGExJs7MCnMnaAWorA4CukNu4ZLKkz2Y kj54apy+/Js= =oHao -----END PGP SIGNATURE-----
"Scott D. Yelich" <[EMAIL PROTECTED]> wrote: > >Um... so is the trigger a new file has been created Yes. >or no more input is >being put into the file? Each one of those begs a question... the first >is, do you want to "tail" or "cat" each and every new file that is >generated Yes. Cyclog keeps a set of at most N log files, each with a preset maximum size. In order to "tail" a cyclog log, you need to tail the most recent file in the log directory. >and the other is "just when do you know that no more input has >been received?" It wouldn't be all that difficult to write a command >that came close to doing those although they seem arbitrarily complicated. I've already posted the program, and it's not all that complicated. -Dave
"Scott D. Yelich" wrote: > > Yup... but now where do I get taildir? Is this more DBWare? Etc. I'm > not sure I know the erason why its necessary to "switch" files or what > the trigger is -- but I'm sure there is a "standard" unix way to > accomplish this. Dave just posted it. See earlier in this thread. > > > The only thing I wish tail would do is multi-tail... mtail... > > > tail -multi, tail file file file... etc. > > You mean monitor several files at once? > > Yes. mtail -f /var/adm/messages /var/log/syslog /bing/bang/boom > /yadda/yadda/yadda /etc/etc/etc. http://www.giovannelli.it/~gmarco/files/xtail.tar.gz Or if you're on FreeBSD: $ cd /usr/ports/misc/xtail ; make install R. -- Robin Bowes - System Development Manager - Room 405A E.O.C., Overseas House, Quay St., Manchester, M3 3HN, UK. Tel: +44 161 838 8321 Fax: +44 161 835 1657
On Thu, 13 May 1999, Scott D. Yelich wrote: Um... so is the trigger a new file has been created or no more input is being put into the file? Each one of those begs a question... the first is, do you want to "tail" or "cat" each and every new file that is generated and the other is "just when do you know that no more input has been received?" It wouldn't be all that difficult to write a command that came close to doing those although they seem arbitrarily complicated. My preference (which is coded in taildir) is that it seek to the end of the newest file upon startup and begin reading there. Thereafter it checks for a new file each time it reads to the end of the current one. If one is found, it is opened and read from the beginning until there is a newer file and so on... Thanks for posting the program, Dave. I really should put up a web page one day. -- Jeff Hayward
-----BEGIN PGP SIGNED MESSAGE----- On 13 May 1999, Robin Bowes wrote: > Dave just posted it. See earlier in this thread. Right. Let me apologize to everyone. I was confused by the seemingly too easy question/problem/issue and I was only trying to help. I'll start a mailing list now where people can chastise me. > > > > The only thing I wish tail would do is multi-tail... mtail... > > > > tail -multi, tail file file file... etc. > > > You mean monitor several files at once? > > Yes. mtail -f /var/adm/messages /var/log/syslog /bing/bang/boom > > /yadda/yadda/yadda /etc/etc/etc. > http://www.giovannelli.it/~gmarco/files/xtail.tar.gz > Or if you're on FreeBSD: > $ cd /usr/ports/misc/xtail ; make install Right. Just with taildir or whatever else package was written for this... I'm sure there is a program out there. I can whip up a little perl proggie to do it as well. I was just saying, apparently not clearly enough, that I just wished that tail would do this *inherently*. Scott ps: has there been any thoughts to making this list moderated? -----BEGIN PGP SIGNATURE----- Version: 2.6.2 iQCVAwUBNzrhUx4PLs9vCOqdAQHGawQA1r77x5MXs9fb0dw/u7UfvHBoLRPwF1Jr yQ5Lo/iwtJGZid78zTET0w9IVXOSfpul9UNrZGVefYyNxkqnmAD3pVl5eiY11RTA MJcKo/pFyAqfhSt0MZfRBxPt2qiIQwecpN/e5IjTQPpYOCyg8jsHrIVmu8/LoYqQ +UQ1kv/VJ6E= =/6w3 -----END PGP SIGNATURE-----
Hmm, the last time I asked this question nobody answered me. so I wrote this: #!/bin/sh QMAIL=`ls /var/log/qmail | grep @ | tail -n 1` QMAILS=`ls /var/log/qmail/qmail-smtpd | grep @ | tail -n 1` tail -f /var/log/daemon.log /var/log/auth.log /var/log/xferlog \ /var/log/mail.log /var/log/kern.log /var/log/cfingerd.log \ /var/log/xntpd /var/log/qmail/$QMAIL /var/log/qmail/qmail-smtpd/$QMAILS I don't have an excessively high volume of mail, and I use rather large cyclog files, so this works well, my syslog logfiles are rotated much more often, causing me to have to break out of this script and restart it anyway. BTW, the tail that ships with most linux's will let you tail -f more than one file. --Adam
From: Scott D. Yelich <[EMAIL PROTECTED]> : Right. Just with taildir or whatever else package was written for : this... I'm sure there is a program out there. I can whip up a little : perl proggie to do it as well. : : I was just saying, apparently not clearly enough, that I just wished : that tail would do this *inherently*. Get GNU tail. SYNOPSIS tail [-c [+]N[bkm]] [-n [+]N] [-fqv] [--bytes=[+]N[bkm]] [--lines=[+]N] [--follow] [--quiet] [--silent] [--verbose] [--help] [--version] [file...] tail [{-,+}Nbcfklmqv] [file...] DESCRIPTION This documentation is no longer being maintained and may be inaccurate or incomplete. The Texinfo documentation is now the authoritative source. This manual page documents the GNU version of tail. tail prints the last part (10 lines by default) of each given file; it reads from standard input if no files are given or when a filename of `-' is encountered. If more than one file is given, it prints a header consisting of the file's name enclosed in `==>' and `<==' before the output for each file.
D. J. Bernstein writes: > DUGRES Hugues, I.T. manager at C.Q.E. writes: > > What do you think of that ??? > > I think that patches are a support nightmare. What you're using isn't > qmail, so don't call it qmail, and don't ask the qmail list for help. I agree. How about giving us permission to distribute a modified form of qmail in source and executable form as long as we call it something other than qmail? If you do that, then the need for patches will go away, and there won't *be* any patches for qmail. -- -russ nelson <[EMAIL PROTECTED]> http://crynwr.com/~nelson Crynwr supports Open Source(tm) Software| PGPok | There is good evidence 521 Pleasant Valley Rd. | +1 315 268 1925 voice | that freedom is the Potsdam, NY 13676-3213 | +1 315 268 9201 FAX | cause of world peace.
On 13 May 1999 05:42:23 -0000, D. J. Bernstein wrote: >I think that patches are a support nightmare. What you're using isn't >qmail, so don't call it qmail, and don't ask the qmail list for help. So, DUGRES Hugues, I.T. manager at C.Q.E., the problem isn't qmail, but in a patch applied to add [desired additional] functionality. What patches are you using in you PIPELINING-violating installation? What can be done to fix the bug in the patch? Which patch should others avoid to avoid violating rfc2197? -Sincerely, Fred (Frederik Lindberg, Infectious Diseases, WashU, St. Louis, MO, USA)
* Ralf Nagel (Mon, May 10, 1999 at 08:19:25PM +0200) > Hi, > > is it possible to handle incoming mail to procmail for further > processing (for example corrections of redundant Reply-to lines or > bad subject lines: the german AW: unstead of Re: etc) and then give > it back to qmail for delivery in a maildir? If all you want is to deliver mail to a maildir, you can use procmail to do this. Use the procmail maildir patch from http://www.vsource.com/~bguenter/distrib/procmail-maildir/ (or from other places, this is the first one I found). Debian's procmail package is already patched. -- SSM - Stig Sandbeck Mathisen Trust the Computer, the Computer is your Friend
Stig Sandbeck Mathisen <[EMAIL PROTECTED]> writes: > If all you want is to deliver mail to a maildir, you can > use procmail to do this. Use the procmail maildir patch... Or, if you'd rather not patch procmail, you can use safecat in recipes under procmail without any patches. You can get safecat at <http://www.pobox.com/~lbudney/linux/software/safecat.html>. -Lenm. -- Scornful men bring a city into a snare: but wise men turn away wrath. --Proverbs 29:8
"Praniti Lakhwara" <[EMAIL PROTECTED] wrote: > >I have a question regarding forwarding. I ahve a mailing address set >up which is supposed to forward to 230 accounts. but when I ad dthe >230 email addresses in its forward to box...and submit it comes back >and says You've got too many unidentified "it"'s there. Be specific: tell us the names of the files and commands you used. >Value for "flags" is too large > >how do I set up teh FLAGS value to take all teh 230 accounts??? Doesn't sound like a qmail problem. I've never heard of "flags". -Dave
Folks, I have a problem with my qmail system. We've been running mail for almost a year now with no major incidents to report.. until now. Our mail queue started filling up yesterday and mail delivery seemed to have slowed down to a crawl.. (It took hours to deliver messages.) I see a very cyclical load average on the machine ranging from .45 to 2.3, up and down every 2-3 minutes or so. The mail queue itself (which is usually around 60-80 megabytes large) has grown to 300 megabytes in the past 24 hours. The machine itself does not seem to be struggling in any way, but the queue just will not flush out. There does not seem to be any (much) spam throughout the mail queue, and there is no steady stream of large messages from any one peticular source. I was hoping that the queue would flush out last night when it was about 250 megabytes, but barely any messages got delivered and it's up to 300 megs this morning. Does anyone have any ideas as to what might be happening on my mail server? Thank you everyone for your help! - Chris Christopher Porreca Systems Administrator N e t A c c e s s , I n c . [EMAIL PROTECTED] _______________________________________________________________ Get Free Email and Do More On The Web. Visit http://www.msn.com
"Christopher Porreca" <[EMAIL PROTECTED]> wrote: > >Does anyone have any ideas as to what might be happening on my mail server? What do your logs say? Trace a single delivery in the logs, and see where the delay is occuring. Run some qmailanalog stats on the logs. What are your concurrencylocal and concurrencyremote settings? Are you seeing that many qmail-local or qmail-remote processes? -Dave
At 05:59 AM 5/13/99 -0700, Christopher Porreca <[EMAIL PROTECTED]> wrote: >Folks, > >I have a problem with my qmail system. We've been running mail for almost a >year now with no major incidents to report.. until now. > >Our mail queue started filling up yesterday and mail delivery seemed to have >slowed down to a crawl.. (It took hours to deliver messages.) I see a very >cyclical load average on the machine ranging from .45 to 2.3, up and down >every 2-3 minutes or so. The mail queue itself (which is usually around >60-80 megabytes large) has grown to 300 megabytes in the past 24 hours. The >machine itself does not seem to be struggling in any way, but the queue just >will not flush out. There does not seem to be any (much) spam throughout >the mail queue, and there is no steady stream of large messages from any one >peticular source. > >I was hoping that the queue would flush out last night when it was about 250 >megabytes, but barely any messages got delivered and it's up to 300 megs >this morning. > >Does anyone have any ideas as to what might be happening on my mail server? >Thank you everyone for your help! What does the maillog file show as reasons for non-delivery? Are the message in the queue (The bulk of them) for local or remote delivery? If remote, are they destined for one particular remote network? Is your DNS server healthy? Gerry
Chris, Look at /var/qmail/lock/trigger It should be a named pipe with the following permissions: prw--w--w- and owned by user:qmails group:qmail When this file gets changed, it can lead to the symptoms you describe. -Jere -- ------------------------------------------------------------------------ // Jere Cassidy - System Administration - D&E SuperNet email: [EMAIL PROTECTED] phone: (717)738-7054 web: http://www.desupernet.net/jere pager/pcs: [EMAIL PROTECTED] - (717)203-0042 ~~~ "While sowing the seeds of Utopia, you invoked a convenient amnesia" -BR ~~~ ------------------------------------------------------------------------ Christopher Porreca wrote: > Folks, > > I have a problem with my qmail system. We've been running mail for almost a > year now with no major incidents to report.. until now. > > Our mail queue started filling up yesterday and mail delivery seemed to have > slowed down to a crawl.. (It took hours to deliver messages.) I see a very > cyclical load average on the machine ranging from .45 to 2.3, up and down > every 2-3 minutes or so. The mail queue itself (which is usually around > 60-80 megabytes large) has grown to 300 megabytes in the past 24 hours. The > machine itself does not seem to be struggling in any way, but the queue just > will not flush out. There does not seem to be any (much) spam throughout > the mail queue, and there is no steady stream of large messages from any one > peticular source. > > I was hoping that the queue would flush out last night when it was about 250 > megabytes, but barely any messages got delivered and it's up to 300 megs > this morning. > > Does anyone have any ideas as to what might be happening on my mail server? > Thank you everyone for your help! > > - Chris > > Christopher Porreca > Systems Administrator > N e t A c c e s s , I n c . > [EMAIL PROTECTED] > > _______________________________________________________________ > Get Free Email and Do More On The Web. Visit http://www.msn.com
Thanks, Dave. Replies below. >What do your logs say? Trace a single delivery in the logs, and see When sending a message it takes some time before it gets reported into my maillog, which was not an issue yesterday. I have also noticed the following from ProcMail: May 13 09:18:19 newman qmail: 926601499.074869 delivery 557041: success: procmai l:_Couldn't_create_"/var/spool/mail/bobb"/did_0+0+1/ >where the delay is occuring. Run some qmailanalog stats on the >logs. What are your concurrencylocal and concurrencyremote settings? >Are you seeing that many qmail-local or qmail-remote processes? I have raised concurrencylocal from 70 -> 150 yesterday, and concurrencyremote from 90 -> 200. I rarely see more than one of each qmail-local and qmail-remote processes running at one time... _______________________________________________________________ Get Free Email and Do More On The Web. Visit http://www.msn.com
We had this problem earlier this year and it turns out that qmail-clean stopped running for some reason. Check and make sure it's still up and running...it cleans the old stuff out of the queue. --Kevin --- Kevin Sawyer - President/CEO - [EMAIL PROTECTED] Applied Personal Computing, Inc. - APCiNet - http://www.apci.net 6001 Old Collinsville Road, Building #3, Fairview Heights, IL 62208 Office: (618) 632-7282 FAX: (618) 632-7287 Support: (618) 628-2Net > -----Original Message----- > From: Christopher Porreca [mailto:[EMAIL PROTECTED]] > Sent: Thursday, May 13, 1999 8:00 AM > To: [EMAIL PROTECTED] > Subject: Mail Queue Just Keeps Growing.... > > > Folks, > > I have a problem with my qmail system. We've been running > mail for almost a > year now with no major incidents to report.. until now. > > Our mail queue started filling up yesterday and mail delivery > seemed to have > slowed down to a crawl.. (It took hours to deliver messages.) > I see a very > cyclical load average on the machine ranging from .45 to 2.3, > up and down > every 2-3 minutes or so. The mail queue itself (which is > usually around > 60-80 megabytes large) has grown to 300 megabytes in the past > 24 hours. The > machine itself does not seem to be struggling in any way, but > the queue just > will not flush out. There does not seem to be any (much) > spam throughout > the mail queue, and there is no steady stream of large > messages from any one > peticular source. > > I was hoping that the queue would flush out last night when > it was about 250 > megabytes, but barely any messages got delivered and it's up > to 300 megs > this morning. > > Does anyone have any ideas as to what might be happening on > my mail server? > Thank you everyone for your help! > > - Chris > > Christopher Porreca > Systems Administrator > N e t A c c e s s , I n c . > [EMAIL PROTECTED] > > > _______________________________________________________________ > Get Free Email and Do More On The Web. Visit http://www.msn.com >
"Christopher Porreca" <[EMAIL PROTECTED]> wrote: > >When sending a message it takes some time before it gets reported into my >maillog, which was not an issue yesterday. Sounds like your "trigger" is hosed. How long is "some time"? >I have also noticed the >following from ProcMail: > >May 13 09:18:19 newman qmail: 926601499.074869 delivery 557041: success: >procmail:_Couldn't_create_"/var/spool/mail/bobb"/did_0+0+1/ You've got two problems here. First is that procmail is trying to deliver to /var/spool/mail. Bobb needs to modify his .procmailrc to deliver to the right place, e.g., tack this onto the end: :0: $HOME/Mailbox The second problem is that procmail doesn't return an exit status that qmail interprets as an error, so messages are going into the bit bucket. I use a procmail wrapper by Phillip Hands, slightly modified for my system: #!/bin/sh # Copyright (c) 1998 Software in the Public Interest <http://www.debian.org/> # Written by Philip Hands <[EMAIL PROTECTED]>. Distributed under the GNU GPL # $Id: qmail-procmail,v 1.2 1998/03/24 19:31:27 phil Exp $ # modified by Dave Sill /var/qmail/bin/preline /usr/local/bin/procmail $* && exit 0 # check if procmail returned EX_TEMPFAIL (75) [ $? = 75 ] && exit 111 # otherwise return a permanent error exit 100 >I have raised concurrencylocal from 70 -> 150 yesterday, and >concurrencyremote from 90 -> 200. I rarely see more than one of each >qmail-local and qmail-remote processes running at one time... If you're not hitting the limit, raising it won't help. :-) -Dave
end +-------------------------------------+ |Greg Albrecht KF4MKT [EMAIL PROTECTED]| |Safari Internet www.safari.net| |Fort Lauderdale, FL 1-888-537-9550| +-------------------------------------+ On Thu, 13 May 1999, Christopher Porreca wrote: >Thanks, Dave. Replies below. > >>What do your logs say? Trace a single delivery in the logs, and see > >When sending a message it takes some time before it gets reported into my >maillog, which was not an issue yesterday. I have also noticed the >following from ProcMail: > >May 13 09:18:19 newman qmail: 926601499.074869 delivery 557041: success: >procmai >l:_Couldn't_create_"/var/spool/mail/bobb"/did_0+0+1/ use: |preline procmail -m /path/to/.procmailrc that works for me. > >>where the delay is occuring. Run some qmailanalog stats on the >>logs. What are your concurrencylocal and concurrencyremote settings? >>Are you seeing that many qmail-local or qmail-remote processes? > >I have raised concurrencylocal from 70 -> 150 yesterday, and >concurrencyremote from 90 -> 200. I rarely see more than one of each >qmail-local and qmail-remote processes running at one time... > > > > > >_______________________________________________________________ >Get Free Email and Do More On The Web. Visit http://www.msn.com >
[EMAIL PROTECTED] wrote: > >Also, I'm wanting to set up qmail so that it's rather unfriendly to people >who telnet straight into it. I want to completely turn off help so that it >doesn't display any version info etc., as well as turning off echo so they >can't see what they're typing. How do I go about this? Nice idea, but there's no way to distinguish between people and programs connected to a port. You could do some heuristics based on the delays between responses, assuming that humans are slower and/or less consistent, but that's not foolproof. Echo isn't handled by the smtp server, it's done by the telnet client, so there's no way to disable it from the server side. -Dave
-----BEGIN PGP SIGNED MESSAGE----- On Thu, 13 May 1999, Dave Sill wrote: > less consistent, but that's not foolproof. Echo isn't handled by the > smtp server, it's done by the telnet client, so there's no way to > disable it from the server side. > -Dave Telnet protocol allows you to turn off echo... but then you're assuming "telnet" protocol. You could probably best do this with a proxy. Scott -----BEGIN PGP SIGNATURE----- Version: 2.6.2 iQCVAwUBNzrPEB4PLs9vCOqdAQHfNgQAyTmtxjE+saPTgePIqOPJG2uggTISTyJG hcK4Pvh+PMCKUS2lhEUA8EXwyxOJlmz1XNnqtcFgAGZSyvDaaCV0HHUHVlZ3ZB7P V5iYlXLfMHymYWpZJ1B0aXaNUAxoE47N5GCpAgywi7T4rsWajYVTQ+f6koRKKkh+ 5BR/fRKz0PM= =KCQt -----END PGP SIGNATURE-----
"Ralf Guenthner" <[EMAIL PROTECTED]> wrote: > >System: SUSE Linux, qmail 1.03 > >Is it possible to set up qmail in such a fashion that it routes >messages for certain recipients, eg. to my address [EMAIL PROTECTED], >not to our normal mail server inside the LAN -whose IP address is in >smtproutes- but directly to another host? I believe you can use fastforward to alias user@host to something else. >Strangely enough qmail already routes a message addressed to >[EMAIL PROTECTED] to the host named mail-gwia... What's strange about that? -Dave
Title: help: tcpserver dies.Hello ,
I have tcpserver running from /var/qmail/rc using the following command line on one line,/usr/local/bin/tcpserver -x/etc/tcp.smtp.cdb -v -u 504 -g 503 0 smtp /var/qmail/bin/qmail-smtpd 2>&1 | /var/qmail/bin/splogger smtpd 3 &
But It keeps dieing
What I usually have to do is run it manually with a NOHUP in front, What am I doing wrong ?Thanks.
-Jhirley
Jhirley Fonte <[EMAIL PROTECTED]> wrote: > > I have tcpserver running from /var/qmail/rc using the following >command line on one line, > >/usr/local/bin/tcpserver -x/etc/tcp.smtp.cdb -v -u 504 -g 503 0 smtp >/var/qmail/bin/qmail-smtpd 2>&1 | /var/qmail/bin/splogger smtpd 3 & > >But It keeps dieing >What I usually have to do is run it manually with a NOHUP in front, >What am I doing wrong ? Take your pick: You're not starting it with "nohup /usr/local/bin/tcpserver ...". You're not running your startup script with csh, bash, or some other shell that disassociates background jobs from the parent. -Dave
On 13 May 1999 06:11:52 -0000, D. J. Bernstein wrote: >If the information changes, there should be a new confirmation message. It may have been read with a different MUA/computer. >list. Trying to cram the same information into the header of every >mesasge is counterproductive. Your solution has the advantage of making it easy to include the subscription address. A simple patch to qmail makes it possible to do that within rfc2369. QMTP/QMQP are no problem since it will be done by the final "exploder". Your solution has the disadvantage of requiring considerable intelligence at the MUA level, a standardized format for confirmation messages, list-id, and new routines at the MLM level; i.e., the likelihood of general support at either end is low. rfc2369 doesn't fix all problems, but all problems can be solved within rfc2369 using qmail [much harder with other MTA]. It's easy to support at both ends and it is an RFC. -Sincerely, Fred (Frederik Lindberg, Infectious Diseases, WashU, St. Louis, MO, USA)
Having the unsubscribe info in the header has saved me a lot of headaches in the last month (ever since I started using the qmail-verh patch). I often have people with a completely messed up envelope sender's address, and they cannot receive a reply to a message sent to the list-help address. I do not understand what better solution Dan is proposing to handle people with bad envelope sender address. ("bad" means, mail cannot be sent there.) People behind fierwalls often have this propblem. Mate
Fred Lindberg writes: > Your solution has the disadvantage of requiring considerable > intelligence at the MUA level, For what? String comparisons on the List-ID? The format is unnecessarily complicated but still manageable: remove spaces and tabs and newlines, remove everything before the last <, squish to lowercase. > a standardized format for confirmation messages, If the MUA is trying to do something fancy like subscribe to a list on the user's behalf, obviously it would appreciate knowing when it has to reply to complete the subscription, and when the user is subscribed. In general, as MUAs learn more and more about mailing lists, there will be more and more demand for easy-to-parse header fields in these messages. > It may have been read with a different MUA/computer. That's a local problem, no more difficult to solve than moving personal address books from one MUA to another. Of course, the problem solves itself when a new confirmation message shows up. ---Dan
On 13 May 1999 19:29:48 -0000, D. J. Bernstein wrote: >For what? String comparisons on the List-ID? The format is unnecessarily No, for finding or archiving confirmation messages. >> a standardized format for confirmation messages, > >If the MUA is trying to do something fancy like subscribe to a list on >the user's behalf, obviously it would appreciate knowing when it has to >reply to complete the subscription, and when the user is subscribed. In >general, as MUAs learn more and more about mailing lists, there will be >more and more demand for easy-to-parse header fields in these messages. It is not trying to do anything on the user's behalf. It is trying to partially hide the problem of different MLM interfaces from the user. The user's [main] problem is how to initiate the desired process without knowing the specific syntax. rfc2369 does that. There is no need for it to be entirely automatic. >> It may have been read with a different MUA/computer. > >That's a local problem, no more difficult to solve than moving personal >address books from one MUA to another. Of course, the problem solves >itself when a new confirmation message shows up. It's not a problem at all if the individual messages contain the necessary info. Assume a subscriber from a new address and MUA mails [EMAIL PROTECTED] What good does the confirmation message do him/her? How can the MLM or MUA determine the actual subscription message without access to a [recent] post from the list? The problem with standards is that they preclude using everybodies best solution in favor of a [lower] common denominator. Let's support rfc2369 as well as possible, while discussing better options. -Sincerely, Fred (Frederik Lindberg, Infectious Diseases, WashU, St. Louis, MO, USA)
Hello All, I would like to know how the qmail book is coming along, and as a side note, I have a domain which is being a PITA in terms of not wanting to stop one of it's customers from auto mailing...the domain is email.com, and I would like to block any mail from being sent from my server and being received by my server...is there an easy way to do this? -Bill
Bill Parker writes: > Hello All, > > I would like to know how the qmail book is coming along, and 3.5 chapters down.... > as a side note, I have a domain which is being a PITA in terms of not > wanting to stop one of it's customers from auto mailing...the domain > is email.com, and I would like to block any mail from being sent from > my server and being received by my server...is there an easy way to > do this? I'd need a better explanation of where their host is and where your host is. -- -russ nelson <[EMAIL PROTECTED]> http://crynwr.com/~nelson Crynwr supports Open Source(tm) Software| PGPok | There is good evidence 521 Pleasant Valley Rd. | +1 315 268 1925 voice | that freedom is the Potsdam, NY 13676-3213 | +1 315 268 9201 FAX | cause of world peace.
Russell Nelson wrote: > Bill Parker writes: > > I would like to know how the qmail book is coming along, and > > 3.5 chapters down.... Howdy, Is the book going to have some content on common patches and/or (hopefully) qmail 2.0? I know there's not much information available on ver2, but I'm thinking of a situation of the book and ver2 becoming available at roughly the same time ... - Tillman Hodgson
Tillman writes: > Russell Nelson wrote: > > > Bill Parker writes: > > > I would like to know how the qmail book is coming along, and > > > > 3.5 chapters down.... > > Howdy, > > Is the book going to have some content on common patches and/or (hopefully) > qmail 2.0? I know there's not much information available on ver2, but I'm > thinking of a situation of the book and ver2 becoming available at roughly the > same time ... I have no inside line to Dan. Writing takes time and book production takes about two months, so if qmail 2.0 comes out this summer, look for the book to be delayed. Given how long his releases have stayed current, I'd really rather write about the finished product anyway. -- -russ nelson <[EMAIL PROTECTED]> http://crynwr.com/~nelson Crynwr supports Open Source(tm) Software| PGPok | There is good evidence 521 Pleasant Valley Rd. | +1 315 268 1925 voice | that freedom is the Potsdam, NY 13676-3213 | +1 315 268 9201 FAX | cause of world peace.
# $MAILPROG = '/var/qmail/bin/qmail-inject'; $MAILPROG = '/usr/sbin/sendmail -t -oeq -n'; I cant' get to work wmail with /usr/sbin/sendmail or /var/qmail/bin/qmail-inject with some simple CGI. Someone can help me ? How can i check if all it's correct ? Best Regards, Luca _ _ ___ | \ / __| Luca Pescatore . Unix Sysadm & Network Manager | \__ \ Network&Solutions Srl . Via Resegone 11 . Arese (Mi) . ITALY |_\_,___/Srl Email: [EMAIL PROTECTED] . Phone: +39-02-9383575
mail:~$ cc taildir.c taildir.c:15: conflicting types for `sys_errlist' /usr/include/errno.h:31: previous declaration of `sys_errlist' taildir.c: In function `newest': taildir.c:65: warning: passing arg 3 of `scandir' from incompatible pointer type taildir.c:65: warning: passing arg 4 of `scandir' from incompatible pointer type am using linux... is there a problem with it? -marlon At 08:46 AM 5/13/99 -0400, Dave Sill wrote: >"Robin Bowes" <[EMAIL PROTECTED]> wrote: >> >>Before I do a bit of coding, has anyone written a script to identify the >>most recent log file and tail it, preferably switching files when the >>log file turns over? > >Jeff Hayward's taildir does what you want. It's small, and I don't >have a URL, so I've attached a copy. > >-Dave > > >Attachment Converted: "g:\abaoma\mail\Attachments\taildir.c" >
On Thu, 13 May 1999, Marlon Anthony Abao wrote: mail:~$ cc taildir.c taildir.c:15: conflicting types for `sys_errlist' /usr/include/errno.h:31: previous declaration of `sys_errlist' taildir.c: In function `newest': taildir.c:65: warning: passing arg 3 of `scandir' from incompatible pointer type taildir.c:65: warning: passing arg 4 of `scandir' from incompatible pointer type am using linux... is there a problem with it? It compiles with 1 warning (which can be ignored) for me. Which Linux are you using? I've got RH 6.0 % uname -v -s -r -m Linux 2.2.5-15 #1 Mon Apr 19 23:00:46 EDT 1999 i686 % cc -o taildir taildir.c taildir.c: In function `newest': taildir.c:65: warning: passing arg 3 of `scandir' from incompatible pointer type % ./taildir . & (sleep 1; echo "Hello, world" >>@0000000002; sleep 3) [2] 1380 Hello, world -- Jeff
It compiled fine for me with no warnings on Debian 2.1. Maybe your headers are outdated. --Adam ----- Original Message ----- From: Jeff Hayward <[EMAIL PROTECTED]> To: <[EMAIL PROTECTED]> Sent: Thursday, May 13, 1999 12:33 PM Subject: Re: taildir won't compile... : On Thu, 13 May 1999, Marlon Anthony Abao wrote: : : mail:~$ cc taildir.c : taildir.c:15: conflicting types for `sys_errlist' : /usr/include/errno.h:31: previous declaration of `sys_errlist' : taildir.c: In function `newest': : taildir.c:65: warning: passing arg 3 of `scandir' from incompatible pointer : type : taildir.c:65: warning: passing arg 4 of `scandir' from incompatible pointer : type : : am using linux... is there a problem with it? : : It compiles with 1 warning (which can be ignored) for me. Which : Linux are you using? I've got RH 6.0 : : % uname -v -s -r -m : Linux 2.2.5-15 #1 Mon Apr 19 23:00:46 EDT 1999 i686 : : % cc -o taildir taildir.c : taildir.c: In function `newest': : taildir.c:65: warning: passing arg 3 of `scandir' from incompatible : pointer type : : % ./taildir . & (sleep 1; echo "Hello, world" >>@0000000002; sleep 3) : [2] 1380 : Hello, world : : : -- Jeff : : :
anyone get it to work on Solaris 2.5? gcc taildir.c Undefined first referenced symbol in file alphasort /var/tmp/cca001nR1.o scandir /var/tmp/cca001nR1.o ld: fatal: Symbol referencing errors. No output written to a.out On Thu, May 13, 1999 at 02:26:47PM -0400, Adam D. McKenna wrote: # It compiled fine for me with no warnings on Debian 2.1. Maybe your headers # are outdated. # # --Adam # # ----- Original Message ----- # From: Jeff Hayward <[EMAIL PROTECTED]> # To: <[EMAIL PROTECTED]> # Sent: Thursday, May 13, 1999 12:33 PM # Subject: Re: taildir won't compile... # # # : On Thu, 13 May 1999, Marlon Anthony Abao wrote: # : # : mail:~$ cc taildir.c # : taildir.c:15: conflicting types for `sys_errlist' # : /usr/include/errno.h:31: previous declaration of `sys_errlist' # : taildir.c: In function `newest': # : taildir.c:65: warning: passing arg 3 of `scandir' from incompatible # pointer # : type # : taildir.c:65: warning: passing arg 4 of `scandir' from incompatible # pointer # : type # : # : am using linux... is there a problem with it? # : # : It compiles with 1 warning (which can be ignored) for me. Which # : Linux are you using? I've got RH 6.0 # : # : % uname -v -s -r -m # : Linux 2.2.5-15 #1 Mon Apr 19 23:00:46 EDT 1999 i686 # : # : % cc -o taildir taildir.c # : taildir.c: In function `newest': # : taildir.c:65: warning: passing arg 3 of `scandir' from incompatible # : pointer type # : # : % ./taildir . & (sleep 1; echo "Hello, world" >>@0000000002; sleep 3) # : [2] 1380 # : Hello, world # : # : # : -- Jeff # : # : # : # # -- /- [EMAIL PROTECTED] --------------- [EMAIL PROTECTED] -\ |Justin Bell NIC:JB3084| Time and rules are changing. | |Pearson | Attention span is quickening. | |Developer | Welcome to the Information Age. | \-------- http://www.superlibrary.com/people/justin/ ----------/
-----BEGIN PGP SIGNED MESSAGE----- On Thu, 13 May 1999, Justin Bell wrote: > anyone get it to work on Solaris 2.5? Not really. I've been told that my system is "seriously" hosed. that would be news to me. I told two people that I wouldn't say this in public... but I just can't keep my big mouth closed. I have accepted the proper and official "answer" that taildir is to be used with cyclog. security [855]> gcc taildir.c Undefined first referenced symbol in file alphasort /var/tmp/cca002d11.o scandir /var/tmp/cca002d11.o ld: fatal: Symbol referencing errors. No output written to a.out security [856]> uname -a SunOS security.spy.org 5.5.1 Generic sun4u sparc SUNW,Ultra-1 narq [146]> gcc taildir.c Undefined first referenced symbol in file alphasort /var/tmp/cclca4Hu1.o scandir /var/tmp/cclca4Hu1.o ld: fatal: Symbol referencing errors. No output written to a.out narq [147]> uname -a SunOS narq.spy.org 5.7 Generic sun4m sparc SUNW,SPARCstation-LX linsux [134]> gcc taildir.c taildir.c: In function `newest': taildir.c:65: warning: passing arg 3 of `scandir' from incompatible pointer type linsux [135]> uname -a Linux linsux.spy.org 2.2.5-15 #1 Mon Apr 19 21:34:49 EDT 1999 sparc unknown I'm not a C genius... but it looks like yet another linux program written without a concern in the world for portibility. linux programs tend to assume that the rest of the world is compatible with them and not the other way around. I'm sure someone could go dig up or translate alphasort and scandir libraries or find compatible calls... but I'm sure if I just RTFM some more or keep searching the FAQs, I'll find the answer there, somewhere. Scott -----BEGIN PGP SIGNATURE----- Version: 2.6.2 iQCVAwUBNzshsx4PLs9vCOqdAQHbGQP+KePfoGhP2eWgJCLG8aqjoaskI3Qq0yoK UGkDumOo7bZZUQYsXYpp//Wg7dzXSAMTiFNIqD2fnHmHHOCEt1hJpP5zQRYKRcr/ qTCigr14eXuNvlJg9taQAgySamaRDDv0SqREpPV3wUm661qAPN8dl5RT9Hvjkv97 sKSQok9JxAo= =qV0G -----END PGP SIGNATURE-----
"Scott D. Yelich" <[EMAIL PROTECTED]> writes: | security [855]> gcc taildir.c | Undefined first referenced | symbol in file | alphasort /var/tmp/cca002d11.o | scandir /var/tmp/cca002d11.o Those functions (as you can tell from the manpage) are part of the BSD compatability library (-lucb), and are not automatically supplied if you just link with the regular solaris C library. Be cautious if you do use libucb, since it interacts very badly with libc sometimes. | I'm not a C genius... but it looks like yet another linux program | written without a concern in the world for portibility. linux programs | tend to assume that the rest of the world is compatible with them and | not the other way around. True, but in this case, it's Solaris that's being revisionist by ditching the traditional Berkeley stuff.
am using slackware 3.6 have finally compiled it by commenting line 15 // extern const char * const sys_errlist[]; and here's what i get: mail:/usr/local/src# uname -a ; cc taildir.c -o taildir; taildir Linux mail1 2.2.7 #3 Wed May 5 21:33:01 PHT 1999 i586 unknown taildir.c: In function `newest': taildir.c:65: warning: passing arg 3 of `scandir' from incompatible pointer type taildir.c:65: warning: passing arg 4 of `scandir' from incompatible pointer type Segmentation fault mail:/usr/local/src# seg faults on me :( -marlon At 02:26 PM 5/13/99 -0400, Adam D. McKenna wrote: >It compiled fine for me with no warnings on Debian 2.1. Maybe your headers >are outdated. > >--Adam > >----- Original Message ----- >From: Jeff Hayward <[EMAIL PROTECTED]> >To: <[EMAIL PROTECTED]> >Sent: Thursday, May 13, 1999 12:33 PM >Subject: Re: taildir won't compile... > > >: On Thu, 13 May 1999, Marlon Anthony Abao wrote: >: >: mail:~$ cc taildir.c >: taildir.c:15: conflicting types for `sys_errlist' >: /usr/include/errno.h:31: previous declaration of `sys_errlist' >: taildir.c: In function `newest': >: taildir.c:65: warning: passing arg 3 of `scandir' from incompatible >pointer >: type >: taildir.c:65: warning: passing arg 4 of `scandir' from incompatible >pointer >: type >: >: am using linux... is there a problem with it? >: >: It compiles with 1 warning (which can be ignored) for me. Which >: Linux are you using? I've got RH 6.0 >: >: % uname -v -s -r -m >: Linux 2.2.5-15 #1 Mon Apr 19 23:00:46 EDT 1999 i686 >: >: % cc -o taildir taildir.c >: taildir.c: In function `newest': >: taildir.c:65: warning: passing arg 3 of `scandir' from incompatible >: pointer type >: >: % ./taildir . & (sleep 1; echo "Hello, world" >>@0000000002; sleep 3) >: [2] 1380 >: Hello, world >: >: >: -- Jeff >: >: >: > > > >
| Segmentation fault | mail:/usr/local/src# | | seg faults on me :( Try running "taildir <name of dir with cyclog files in it>" instead of just taildir. Troy
-----BEGIN PGP SIGNED MESSAGE----- On Thu, 13 May 1999, Troy Morrison wrote: > | Segmentation fault > | mail:/usr/local/src# > | seg faults on me :( > Try running "taildir <name of dir with cyclog files in it>" instead of > just taildir. > Troy solaris 2.5.1 security [965]> gcc -L/usr/ucblib -lucb taildir.c -o taildir security [966]> ./taildir . ./taildir: scandir: : No such file or directory I also can't seem to get a -static version of this to compile... and I had to add /usr/ucblib to my LD_LIBRARY_PATH. but, as you can see, it still doesn't run. Yes, it's a solaris issue where they don't have th bsd stuff... but, alas, sunos was bsdish and solaris is sysvish.... I'm actually impressed by just how much bsd stuff there still is in solaris. Anyway, taildir still doesn't "work" ... although it mostly compiles and stuff now. What's the next step? Scott -----BEGIN PGP SIGNATURE----- Version: 2.6.2 iQCVAwUBNzuFIx4PLs9vCOqdAQHKbgP+Kqhcohbbr046ZIkS5Xm3gt22Xhu64ru0 rW0k76ePyh7VmPMwqfYNlwdV5KRYtP9txiEccP45fH0bR/Kw1gSwogzseNrg058B mUtIYpQyXxE1bIV+KmtxcdtW7amUSJb67hKTbYmX9hmscEjg9wIoQ1Y1ut2sUkrT 2tWS67r5NyM= =+Bki -----END PGP SIGNATURE-----
"Scott D. Yelich" wrote: > > security [965]> gcc -L/usr/ucblib -lucb taildir.c -o taildir > security [966]> ./taildir . > /taildir: scandir: : No such file or directory > > I also can't seem to get a -static version of this to compile... > and I had to add /usr/ucblib to my LD_LIBRARY_PATH. > > but, as you can see, it still doesn't run. Try using are ful directory name rather than ".", eg: $ ./taildir /var/log/qmail Just a thought... -- Two rules to success in life: 1. Don't tell people everything you know. -- Sassan Tat
Hi, OK i've posted a few messages to the list thanks for those who responded, i'm still at point blank. With my current qmail setup (without modifing qmail boot scripts) everything queues to /var/qmail/autoturn/ip/new correctly for all the relay domains (theres about 50 directories already queued). When I modifiy the script to the following which as far as i know is correct: /usr/bin/tcpserver -u 71 -g 65534 0 smtp \ sh -c ' /usr/sbin/qmail-smtpd cd /var/qmail/autoturn exec /usr/bin/setlock -nx $TCPREMOTEIP/seriallock \ /usr/bin/maildirsmtp $TCPREMOTEIP autoturn-$TCPREMOTEIP- $TCPREMOTEIP AutoTURN' & tcpserver starts correctly and binds to the smtp port, if i telnet in smtp is funtioning correctly. If I send a test mail it is queued properly as it should be. For some reason it won't send the client there mail!!! Its all queued just won't send it on. Please could someone help me get this sorted, any tips or advice with be greatly apprieated. Qmail is a great product the only thing I do find is the documentation does not really help you get your answer. Thanks, Chris.
Our server is MX for 3 different domains. Is there any way to log qmail info in 3 different log files? Is there any way to log this info (with syslog.conf) to even 3 different hosts? G.
On Thu, May 13, 1999 at 05:59:28PM +0300, Stathakopoulos Giorgos wrote: > Our server is MX for 3 different domains. Is there any way to log qmail > info in 3 different log files? Is there any way to log this info (with > syslog.conf) to even 3 different hosts? 1. Run 3 different qmail, and have them log to different log files using cyclog. 2. Pass qmail's logs to a shell script which can then send the log to syslog at different facilities (eg. local1, local2, etc). Then configure your syslog.conf to log these different facilities in different files/hosts. 3. Why do you need to split the logs? You can use qmailananlog to analyse the logs and extract all relevant information about your 3 domains. 4. I've run out of ideas! -- System Administrator See complete headers for address, homepage and phone numbers
After implementing the new scripts for Procmail and completely restarting the qmail daemons, qmail has now flushed 150 megs of queue in 45 minutes! Thank you everyone for your assistance in this emergency. I did discover that procmail (even though we're not running sendmail in any way) still requires the /var/spool/mail/ directory for temporary lock files. By bringing down my concurrencyremote to 1 session, this is also helping force the queue to flush down to all of my local users. Also thanks for the procmail filter script that returns the correct errors, this has helped astronomically. - Chris Christopher Porreca Systems Administration N e t A c c e s s , I n c . [EMAIL PROTECTED] (Which should work again, now!) _______________________________________________________________ Get Free Email and Do More On The Web. Visit http://www.msn.com
On Wed, May 12, 1999 at 04:53:00PM +0200, Pavel Kankovsky wrote: > no comment I had one similar to this one the other day. That one involved MDaemon, another Windows SMTP product. Greetz, Peter -- | 'He broke my heart, | Peter van Dijk | I broke his neck' | [EMAIL PROTECTED] | nognikz - As the sun | Hardbeat@ircnet - #cistron/#linux.nl | | Hardbeat@undernet - #groningen/#kinkfm/#vdh |
I got this from a friend on the ZMailer list.. ----- Forwarded message from Matti Aarnio <[EMAIL PROTECTED]> ----- Subject: Re: Zmailer and Qmail are not fully compatible ... From: Matti Aarnio <[EMAIL PROTECTED]> To: [EMAIL PROTECTED] (Matti Aarnio) Date: Wed, 12 May 1999 19:06:20 +0300 (EET DST) Cc: [EMAIL PROTECTED], [EMAIL PROTECTED], [EMAIL PROTECTED] (following on my previous letter topics) After a bit more of manual testing: ... > That is "250 ok" for the MAIL, but no replies to "RCPT" or "DATA". > It looks like PIPELINING does not work at the QMAIL after all ? > Then your quick cure would be to remove the "PIPELINING" text > from the smtp server's EHLO responses. > > Timeout (300 sec) while waiting responses from remote > > Huh ? And ZMailer proceeds with message sending ? > Now that might be a bug here, indeed.. (and produce the > "hundreds of '500' errors you mention") Yes, that it definitely is, now I have a sort of fix for it, too. (but it isn't complete, I will return to that.) The test seem to indicate that QMAIL does violate PIPELINING rule (see RFC 2197) that says: (chapter 4.2): (1) MUST NOT flush or otherwise lose the contents of the TCP input buffer under any circumstances whatsoever. This could be said more clearly. Like: "When reading inputs from TCP input stream, it shall consume only those characters that the command line currently under processing does contain, and NEVER loose data related to latter lines in the input, which may appear in lowlevel socket read function." At ZMailer this is done with double-buffering, socket is read into a buffer from which a local special version of getc() (and gets() calling it) does reading, and if needed, buffer refilling. What I do find surprising is that with PIPELINING input of: MAIL From:<[EMAIL PROTECTED]> BODY=8BITMIME RCPT To:<[EMAIL PROTECTED]> (pausing here for the replies) works just fine, but: MAIL From:<[EMAIL PROTECTED]> BODY=8BITMIME RCPT To:<[EMAIL PROTECTED]> RCPT To:<[EMAIL PROTECTED]> RCPT To:<[EMAIL PROTECTED]> RCPT To:<[EMAIL PROTECTED]> DATA Yields only two replies: 250 ok 250 ok waiting some 20-30 seconds, I can issue: RSET 250 flushed and it is honoured. That tells me that your instance of QMAIL does break PIPELINING rules at the RCPT processing. A quick cure for you is to use some binary editing capable editor e.g. emacs, and replace the: 250-PIPELINING\r\n with e.g.: 250-X-PELINING\r\n Then systems capable to use PIPELINING when sending to your smtp server will cease of doing so. /Matti Aarnio <[EMAIL PROTECTED]> ----- End forwarded message ----- ----- End forwarded message ----- Greetz, Peter -- | 'He broke my heart, | Peter van Dijk | I broke his neck' | [EMAIL PROTECTED] | nognikz - As the sun | Hardbeat@ircnet - #cistron/#linux.nl | | Hardbeat@undernet - #groningen/#kinkfm/#vdh |
Hi, I have a client who attempts to email email addresses (he says any address he tries) and gets the following responses: The message could not be sent because one of the recipients was rejected by the server. The rejected e-mail address was ' <mailto:'[EMAIL PROTECTED]'> [EMAIL PROTECTED]'. Subject 'Test about something', Account: 'test2.test2.com.au', Server: 'test2.test2.com.au', Protocol: SMTP, Server Response: '553 sorry, that domain isn't in my list of allowed rcpthosts (#5.7.1)', Port: 25, Secure(SSL): No, Server Error: 553, Error Number: 0x800CCC79 My rcpthosts file just contains: [root@rhino control]# cat rcpthosts localhost test2.test2.com.au test3.com I've just changed the actual domains above from their real alternatives for this list. I have no idea why this is happening, does anyone have any suggestions? I appreciate your help. Michael.
How do I get off this list? Dirk On Fri, May 14, 1999 at 11:23:41AM +1000, Michael Mansour wrote: > Hi, > > I have a client who attempts to email email addresses (he says any address > he tries) and gets the following responses: > > The message could not be sent because one of the recipients was rejected by > the server. The rejected e-mail address was ' > <mailto:'[EMAIL PROTECTED]'> [EMAIL PROTECTED]'. Subject 'Test > about something', Account: 'test2.test2.com.au', Server: > 'test2.test2.com.au', Protocol: SMTP, Server Response: '553 sorry, that > domain isn't in my list of allowed rcpthosts (#5.7.1)', Port: 25, > Secure(SSL): No, Server Error: 553, Error Number: 0x800CCC79 > > My rcpthosts file just contains: > > [root@rhino control]# cat rcpthosts > localhost > test2.test2.com.au > test3.com > > I've just changed the actual domains above from their real alternatives for > this list. I have no idea why this is happening, does anyone have any > suggestions? > > I appreciate your help. > > Michael.
On Fri, May 14, 1999 at 11:23:41AM +1000, Michael Mansour wrote: > I have a client who attempts to email email addresses (he says any address > he tries) and gets the following responses: > > The message could not be sent because one of the recipients was rejected by > the server. The rejected e-mail address was ' > <mailto:'[EMAIL PROTECTED]'> [EMAIL PROTECTED]'. Subject 'Test > about something', Account: 'test2.test2.com.au', Server: > 'test2.test2.com.au', Protocol: SMTP, Server Response: '553 sorry, that > domain isn't in my list of allowed rcpthosts (#5.7.1)', Port: 25, > Secure(SSL): No, Server Error: 553, Error Number: 0x800CCC79 You need to read a bit of the documentation, in particular FAQ 5.4 and ftp://koobera.math.uic.edu/www/qmail/faq/servers.html#authorized-relay, which is from the updated FAQ. For more detail on the subject, read http://www.palomine.net/qmail/relaying.html and http://www.palomine.net/qmail/selectiverelay.html Chris
[EMAIL PROTECTED] writes: | How do I get off this list? Good question. Maybe the list manager should put that information at the bottom of each and every message, so you won't have to remember what it told you back when you first subscribed. What say, Dan?
Hi All, Im using QMail under Debian 2.1 with Autoturn to provide an SMTP relayer for dialup SMTP Services. Everything is working on inbound delivery, however when a user tries to send mail through the relay, it is denied with an error: 553 sorry, that domain isnt in my list of allowed rcpthosts (#5.7.1) I have a rcpthosts file containing a list of the domains which are allowed to relay through the server, and also the list of IP's listed for tcpserver to allow. This is now becoming an urgent matter as we are failing to provide customers with a reliable service, if someone out there knows the answer (which im sure is simple), please let me know. Regards Gavin -- Gavin Lewandowski - 1Way Internet Ltd - Connecting the West to the World.... Network Systems Administrator T: +44 117 9414141 F: +44 117 9413430 DDI: +44 117 9392040 ICQ: 1702997 ORANGE: +44 7970 728645 SMS: [EMAIL PROTECTED] FINGER: [EMAIL PROTECTED] SNAIL: Olympia House, Beaconsfield Road, St. George, Bristol, BS5 8ER
On Fri, May 14, 1999 at 10:34:05AM +0100, Gavin Lewandowski wrote: http://qmail-docs.surfdirect.com.au/docs/qmail-antirelay.html > Hi All, > > Im using QMail under Debian 2.1 with Autoturn to provide an SMTP relayer > for dialup SMTP Services. > > Everything is working on inbound delivery, however when a user tries to > send mail through the relay, it is denied with an error: > > 553 sorry, that domain isnt in my list of allowed rcpthosts (#5.7.1) > > I have a rcpthosts file containing a list of the domains which are allowed > to relay through the server, and also the list of IP's listed for > tcpserver to allow. > > This is now becoming an urgent matter as we are failing to provide > customers with a reliable service, if someone out there knows the answer > (which im sure is simple), please let me know. -- System Administrator See complete headers for address, homepage and phone numbers
