Re: Is qmail best reserved for mailing list server purposes only?
Russ Allbery wrote: Rather, it tries to bounce them and the bounce bounces as undeliverable. The solution is for ORBS to stop probing systems from which no spam has ever been sent and for which there is no reason to suspect a lack of security. they were a lot easier to igore when they were still calling themselves dorkslayers -- David Nicol 816.235.1187 [EMAIL PROTECTED] Parse, munge, repeat.
Re: Is qmail best reserved for mailing list server purposes only?
Oleg Polyakov wrote: I'm not sure how qmail works if you are sending 100 messages from server to another one. Does it open 100 connections concurrently? it opens maxconcurrency connections. It doesn't have per-site concurrency limit, unles you patch it. It is reccommended, if you are having a problem killing a particular smtp peer, to trap all outgoing mail for it by defining it as a local virtual host, and then passing the stack of mail in the local virtual host's MailDir to the peer with something called serialmail. -- David Nicol 816.235.1187 [EMAIL PROTECTED] Parse, munge, repeat.
Re: Pop goes my weasel
Meuse, Andy wrote: Hey all, A few accounts on my qmail server recieve 1000 emails a day. Sometimes these don't get checked for weeks. The mail is also kept on the server for a few weeks so the CUR dir gets pretty massive. It's all been running fine for months with no config change, but now when a user tries to pop the account the cpu% on the server maxes out. If the user quits Outlook and then starts again, a second pop process starts on the qmail server and the cpu% splits between them. If I kill the processes it all just starts again. The only way around it I've found is to delete messages from the CUR dir down to about 1000 or so. The server is a dual 500 with half a gig of ram and the desktops are 700s with 256 and the connectivity is not an issue. Anyone know of some tweak that might help me? Or do y'all need more info? Thanks Andy I rewrote my pop daemon to only serve 200 messages at once. When there more messages than that I need to make several requests. You might experiment with some internal scaling tool that makes a new directory every few thousand mails and moves all the mail from the MailDir/new directory into the newly created directory, and a pop daemon (again, assuming you are writing your own pop daemon) that knows about this multiple directory arrangement and naming scheme (new+timestamp would work well) That way you could avoid the too-many-entries-in-directory problem tht you appear to be having. -- David Nicol 816.235.1187 [EMAIL PROTECTED] Described as awesome by users
Re: X-Sender
[EMAIL PROTECTED] wrote: Hi, Which qmail process write the X-Sender field in the headers? I would like to remove it, or to rewrite it, because it uses the name of my host/domain, which are not real. BTW, how can i rewrite any header of my outgoing mail? David Gmez "The question of whether computers can think is just like the question of whether submarines can swim." -- Edsger W. Dijkstra I wrote a patch to have qmail-smtpd add the hostname to an incomplete Sender: header you could do something similar to leave the sender field the hell out. My patches are at http://davidnicol.com/qmail.html Yes, it's true. But another idea of mine emerges - is there a way (in qmail) to _globally_ add my custom header to all outgoing mail ? Cheers, -- Lukasz Felsztukier Edit your qmail-remote program to stick it in there -- David Nicol 816.235.1187 [EMAIL PROTECTED] He who says it's impossible shouldn't interrupt the one doing it.
Re: Clustering qmail servers
Tracy R Reed wrote: On Fri, Feb 09, 2001 at 01:08:47PM -0600, Herbie wrote: Well the simplest way is to have one machine act as the gateway for all mail and create alias files to forward the mail onto the second machine. I used a simple perl script from a flat file to create the .qmail alias's. I guess that could work but there is no easy automated way to manage so many qmail files and we already have 1760 in there already. I think I'll just have my qmail-queue wrapper rewrite the envelope recipient address and add a headerline which is basically what qmail-alias does when it forwards an email on somewhere else. I was just wondering if anyone came up with a more correct solution but it seems not. I'd think a NFS solution would be appropriate, so the SMTP boxes and the POP boxes can all be different boxes, that access the same user directories. This is the whole point of maintaining MailDir NFS-safety isn't it? -- David Nicol 816.235.1187 [EMAIL PROTECTED] "Nothing in the definition of the word `word' says that a word has to be in a dictionary to be called one." -- Anu Garg
Re: Load Balancing
I have a server called MLM and 4 servers called
MLM1,2,3,4
.
MLM is a central server with Qmail and EZMLM, and the other servers are
the RELAY
Run this to start your load balancing:
perl -e'chdir"/var/control/";while(1){sleep(1);system "echo :MLM${\(++$n%4 + 1)}sr_";
rename "sr_","smtproutes"}'
you might want to change the sleep interval to something higher.
Or change your relays to MLM0,1,2,3 so you can leave out the +1
Multi-level marketing sucks though
--
David Nicol 816.235.1187 [EMAIL PROTECTED]
"I don't care how they do it in New York"
QMTP protocol spec question
the QMTP spec includes: 8. Examples A client opens a connection and sends the concatenation of the following strings: "246:" 0a "Received: (qmail-queue invoked by uid 0);" " 29 Jul 1996 09:36:40 -" 0a "Date: 29 Jul 1996 11:35:35 -" 0a "Message-ID: [EMAIL PROTECTED]" 0a "From: [EMAIL PROTECTED]" 0a "To: [EMAIL PROTECTED] (D. J. Bernstein)" 0a 0a "This is a test." 0a "," "24:" "[EMAIL PROTECTED]" "," "30:" "26:[EMAIL PROTECTED]," "," "356:" 0d "From: [EMAIL PROTECTED]" 0d 0a "To:" 0d 0a " Hate." 22 "The Quoting" 22 "@SILVERTON.berkeley.edu," 0d 0a " " 22 "\\Backslashes!" 22 "@silverton.BERKELEY.edu" 0d 0a 0d 0a "The recipient addresses here could" " have been encoded in SMTP as" 0d 0a "" 0d 0a " RCPT TO:Hate.The\ [EMAIL PROTECTED]" 0d 0a " RCPT TO:\\[EMAIL PROTECTED]" 0d 0a 0d 0a "This ends with a partial last line, right here" "," "0:" "," "83:" "39:Hate.The [EMAIL PROTECTED]," "36:\[EMAIL PROTECTED]," "," The server sends the following response, indicating acceptance: "21:Kok 838640135 qp 1390," "21:Kok 838640135 qp 1391," "21:Kok 838640135 qp 1391," The client closes the connection. I am confused. Why are there three responses for two recip. addrs? The zero length is the envelope sender -- Is that acceptable? Is the is the second server response doubled in the http://cr.yp.to/proto/qmtp.txt document? -- David Nicol 816.235.1187 [EMAIL PROTECTED] "gorkulator borked. Please investigate."
Re: Using a RAMDISK for /var/qmail/queue thoughts ?
David Dyer-Bennet wrote: Um, most reporting measured results from optimizing high-traffic qmail-based mail servers have found that disk activity on the queue disk is the first limit they hit. How about, if the first delivery fails, pass it off to a server with some disks. Why not pre-process with qmail-remote before queueing? -- David Nicol 816.235.1187 [EMAIL PROTECTED] Five seconds of light is a lot of data.
Re: A firestorm of protest?
Chris Garrigues wrote: "Upgrade" suggests adding features, rather more than "patch" does; patches are often released to fix bugs. How about "addition" or "extension"? we need something that vaguely impugns the patch, without implying that the patch is required, and we wish to keep current meaning of "patch" and be consistent with all current habits. My nomination is, drumroll please "non-standard option" or , even more impugnly, "unsupported option" These could even be ranked in order of sanity, from the ones that get mentioned all the time on the list, to the ones that are heretical to "official" reccommendations. -- David Nicol 816.235.1187 [EMAIL PROTECTED] "Live fast, die young, and leave a beautiful corpse"
Re: Dot in email adress
James R Grinter wrote: "David L. Nicol" [EMAIL PROTECTED] writes: What exactly is the threat this is supposed to guard against? Is it directory descending on vms, or access to the .. directory somehow? I think it's along the lines of something like 'user-/../foo@domain' which would naively search for '.qmail-/../foo'. Replacing '.' is an easy way to prevent it ever being possible. James. Yes, but dot appears so many places -- would not replacing slash be a better solution? -- David Nicol 816.235.1187 [EMAIL PROTECTED] "people with fish eyes and brown socks"
Re: Dot in email adress
Johan Almqvist wrote: man 5 dot-qmail replace the dot (.) with a colon (:) in the name of the .qmail file, ie .qmail-ar:rubin -Johan that man page says: WARNING: For security, qmail-local replaces any dots in ext with colons before checking .qmail-ext. For convenience, qmail-local converts any uppercase letters in ext to lowercase. What exactly is the threat this is supposed to guard against? Is it directory descending on vms, or access to the .. directory somehow? I am not aware of a special case where dots in a file name will unexpectedly hork a unix file system -- is it an obsolete fear, or a current one that I don't know about? This seems, without knowing what the threat is, an arbitrary exception. -- David Nicol 816.235.1187 [EMAIL PROTECTED] "people with fish eyes and brown socks"
fifo smtproutes Re: Qmail and Large Scale Dynamic Mailings
what if /var/qmail/control/smtproutes was replaced with a fifo that
gave a different relay every time it was read?
#!/usr/local/bin/perl
while(++$count){
unlink '/var/qmail/control/smtproutes';
system 'mkfifo /var/qmail/control/smtproutes';
$c=$count % 5; # or however many there are
open R,"/var/qmail/control/smtproutes"; #block until it is read
print R ":bsdrelay$c.macrosys.com\n";
};
__END__
Will the above cause unexpected freezes?
A less intensive solution might be to overwrite the tenth
character in the static file every few seconds, to load up that
relay.
"Collin B. McClendon" wrote:
Hello,
Sounds good.
Thanks,
Collin
-Original Message-
From: David L. Nicol [mailto:[EMAIL PROTECTED]]
Sent: Thursday, January 11, 2001 9:58 AM
To: Collin B. McClendon
Subject: Re: Qmail and Large Scale Dynamic Mailings
several slave BSD boxes with high concurrencies and a hacked qmail-remote
that round-robins through them.
--
David Nicol 816.235.1187 [EMAIL PROTECTED]
"people with fish eyes and brown socks"
Re: emulating sendmail's user@host.REDIRECT feature?
Matt Harrington wrote: Great! that does it. Any idea how to include a newline in the error though? along the lines of... | bouncesaying '\nMy new address is:\n\[EMAIL PROTECTED]' ---Matt how about | bouncesaying [EMAIL PROTECTED] echo My new address is More than that, opening up bouncesaying.c and adding your own verbiage should not be too difficult; or making any other program that exits 100. Does that work, or does the bounce protocol only have room for a one-line reason? -- David Nicol 816.235.1187 [EMAIL PROTECTED] "the toad doesn't know..."
Re: footer patch??
Johan Almqvist wrote:
On Mon, Jan 08, 2001 at 10:00:24PM +0100, BOFH wrote:
is there any patch, which add some text at the bottom of each sending
message?
It can be done using the QMAILQUEUE patch by Bruce G - but it's dangerous
as it may break MIME. Or complicated, as you'll have to parse the MIME
message, add the footer and then en-MIME it again.
-Johan
if you don't mind jamming it on at end of MIME,you could open up qmail-remote
and stick it within blast(), like this:
char *Footer = "\r\nAlmqvist Industries makes no claims of\r\n"
"the accurracy of any claims made by any of\r\n"
"our employees.\r\n"; /* new */
void blast()
{
int r;
char ch;
for (;;) {
r = substdio_get(ssin,ch,1);
if (r == 0) break;
if (r == -1) temp_read();
if (ch == '.')
substdio_put(smtpto,".",1);
while (ch != '\n') {
substdio_put(smtpto,ch,1);
r = substdio_get(ssin,ch,1);
if (r == 0) perm_partialline();
if (r == -1) temp_read();
}
substdio_put(smtpto,"\r\n",2);
}
substdio_put(smtpto,Footer,strlen(Footer),102); /* new */
flagcritical = 1;
substdio_put(smtpto,".\r\n",3);
substdio_flush(smtpto);
}
Does that work?
--
David Nicol 816.235.1187 [EMAIL PROTECTED]
"Warning: may contain incomprehensible strings of random letters"
Re: Backup Qmail Server
[EMAIL PROTECTED] wrote: On Wed, 27 Dec 2000, Michael Hornby wrote: My ultimate goal is to have my home server accept any mail that is being sent to any e-mail address being hosted on the main server, and to indefinitely try to forward it to the main server. This way, when the main server returns, it will receive all the mail it missed while it was down. All the e-mail will then continue to be stored on the main server, and users can login there to retrieve it. the only extra thing you have to do is make sure example.com is in rcpthosts on backup.example.net and also does not appear in locals or virtualdomains on that machine. (if it does then you have to do htings slightly differently) RjL There's also rigging something to kill -ALRM the home server when the main server wakes up, if it has been down a while; as long as that won't smother it, setting concurrencyremote on home low should help. -- David Nicol 816.235.1187 [EMAIL PROTECTED] "the toad doesn't know..."
Re: Attachment-based relaying
Brett Randall wrote: how to intercept mail that our users send through our mail server, check the size of the mail, and if it exceeds a certain size (say, 5mb), then it relays the mail to another qmail relay, otherwise the current relay treats it as normal outgoing e-mail. Does anyone have ideas as to how I would implement this? TIA you could alter your qmail-remote, to connect to your peer that uses the alternate interface instead of the proper destination, when the size is too big. Sort of like the hack to defer when a message is too big but instead of deferring the peer-address is changed. For a starting point here is how to find the file size of the message within qmail-remote.c #include sys/stat.h int instat; fstat( ssin.fd, instat); -- David Nicol 816.235.1187 [EMAIL PROTECTED] Today in art class, draw your sword
Re: Qmail and Large Scale Dynamic Mailings
if the newsletters are all the same, you could pre-process the list, organizing by recipent domain and starting qmail-remote processes with a few dozen recipients, and only give qmail the ones that don't go through on the first attempt, and even then after some back-off time. That way you won't clog up all incoming channels somewhere, denying service, and getting barred. Some perl code that can work with qmail-remote is available at http://www.davidnicol.com/qmail.html You could open one channel per recipent domain by inserting fork and next; right after the for(keys %Recipients) line. Adding a random backoff timer, something like sleep(rand 3600) before invoking qmail-inject might ease things too, as long as your machine has enough swap space to have all these forked perls sitting around sleeping. YMMV. Henning Brauer wrote: Am Mittwoch, 6. Dezember 2000 22:04 schrieb Thomas Duterme: How about increasing your concurrencyremote to something like 100? you most likely are hitting your limits. Good point. Will try that tonight. I've gotten some problems before from ISP's blocking us when I went up to 240...I'm not quite sure what the highest polite limit on this should be. Hmm, even with 20 concurrent connections our servers was blocked by some braindead freemailer's servers when one of our customers sent out a newsletter... I don't think there is a common "highest polite limit", you have to figure it out for your country, even for your typical recipients. Greetings Henning -- Henning Brauer | BS Web Services Hostmaster BSWS| Roedingsmarkt 14 [EMAIL PROTECTED] | 20459 Hamburg www.bsws.de| Germany -- David Nicol 816.235.1187 [EMAIL PROTECTED] idamfino/i
patch to be kind to broken MUAs that do not include host name on a sender line
a pathologially selective listserv that needs to be replaced
brought to my attention the fact that my MUA has been inserting
a
Sender: david
line in the headers of my messages, out of accordance with rfc 850
which apparently specifies that the host-name is supposed to go there.
Rather than change MUA, I have modified qmail-smtpd.c to add
@remotehost
to incomplete sender lines it might receive.
Here's what I did:
***
*** 345,350
--- 348,380
qmail_put(qqt,ch,1);
}
+ void TweakSendera(pch)
+ char *pch;
+ { /* add @host to sender header if not provided */
+char *qch;
+
+ qch = "(ABBAZA)";
+ while (*qch){ qmail_put(qqt,qch++,1); }
+
+
+ }
+ void TweakSender(pch)
+ char *pch;
+ { /* add @host to sender header if not provided */
+char ch;
+char *qch;
+
+for(ch = *pch; ch != '\r' ;substdio_get(ssin,ch,1)){
+ if (ch == '@') {*pch = ch; return; }/* no tweaking required */
+ put(ch);
+} /* finishing the loop means \r was encountered before @ */
+ *pch = ch;
+ qmail_put(qqt,"@",1);
+ qch = remotehost;
+ while (*qch){ qmail_put(qqt,qch++,1); }
+ }
+
+
void blast(hops)
int *hops;
{
***
*** 355,361
int flagmaybex; /* 1 if this line might match RECEIVED, if fih */
int flagmaybey; /* 1 if this line might match \r\n, if fih */
int flagmaybez; /* 1 if this line might match DELIVERED, if fih */
!
state = 1;
*hops = 0;
flaginheader = 1;
--- 385,392
int flagmaybex; /* 1 if this line might match RECEIVED, if fih */
int flagmaybey; /* 1 if this line might match \r\n, if fih */
int flagmaybez; /* 1 if this line might match DELIVERED, if fih */
! int flagmaybes; /* 1 if this line might match SENDER, if fih */
!
state = 1;
*hops = 0;
flaginheader = 1;
***
*** 369,379
if (pos 8)
if (ch != "received"[pos]) if (ch != "RECEIVED"[pos]) flagmaybex = 0;
if (flagmaybex) if (pos == 7) ++*hops;
if (pos 2) if (ch != "\r\n"[pos]) flagmaybey = 0;
if (flagmaybey) if (pos == 1) flaginheader = 0;
}
++pos;
! if (ch == '\n') { pos = 0; flagmaybex = flagmaybey = flagmaybez = 1; }
}
switch(state) {
case 0:
--- 400,413
if (pos 8)
if (ch != "received"[pos]) if (ch != "RECEIVED"[pos]) flagmaybex = 0;
if (flagmaybex) if (pos == 7) ++*hops;
+ if (pos 6)
+ if (ch != "sender"[pos]) if (ch != "SENDER"[pos]) flagmaybes = 0;
+ if (flagmaybes) if (pos == 6) TweakSender(ch);
if (pos 2) if (ch != "\r\n"[pos]) flagmaybey = 0;
if (flagmaybey) if (pos == 1) flaginheader = 0;
}
++pos;
! if (ch == '\n') { pos = 0; flagmaybes = flagmaybex = flagmaybey = flagmaybez =
1; }
}
switch(state) {
case 0:
--
David Nicol 816.235.1187 [EMAIL PROTECTED]
Re: secrets and lies
Ian Lance Taylor wrote: Date: Wed, 29 Nov 2000 18:34:59 -0800 From: Greg White [EMAIL PROTECTED] I can't see any circumstances where any of Dan's sofware can be deemed closed source. It is not the case that all software is either open source or closed source. There is a broad continuum of licensing possibilities. I already mentioned an important freedom which Dan does not permit. The lack of that freedom means that Dan's software is not open source. Saying that Dan's software is not open source does not mean that it is closed source. Dan's software is almost open source, it just isn't quite all the way there. Ian http://courier.sourceforge.net/ appears to be a GPL'd qmail clone, more or less. Why not use it instead, you want a GPL MTA? -- David Nicol 816.235.1187 [EMAIL PROTECTED] Just when you think you're finally safe, the poets reappear
Re: 1.04---not
Felix von Leitner wrote: At least: has anybody thought about implementing MXPS: http://cr.yp.to/proto/mxps.txt Several people have. But it is not worth the bother until a noticable part of the Internet uses it. Felix What is the advantage of MXPS over SMTP options? It seems like the SMTP option framework is flexible enough to do anything with, for instance encryption or compression, within the confines of your connection. -- David Nicol 816.235.1187 [EMAIL PROTECTED] Just when you think you're finally safe, the poets reappear
Re: secrets and lies
Instead, it poses the question: do you have the legal right to use the web, in the absence of explicit copyright notices on every document element you encounter? Laws are never about what is allowed. Laws are about what is prohibited.
Re: return receipts
Yes, exactly. Without doubt the behavior of this typically underhyped feature, like the rest of the dot-qmail file system, depends on which .qmail file is selected, for fine control over which of the various inboxes will receive delivery notifications and which e-mail address will appear to be the deliveree. Andy Bradford wrote: Thus said "David L. Nicol" on Mon, 20 Nov 2000 20:12:46 CST: What about the "notification on delivery" stuff -- is that not an MTA feature? Is it deprecated? Rather it would be a feature of the MDA, has anyone added it to qmail-local? You mean something like what is covered in "man qreceipt" ? It depends on what the user is expecting I guess... If it is Disposition-Notification-To then it has nothing to do with the MTA, however, if it is Notice-Requested-Upon-Delivery-To then that is covered... Andy -- [---[system uptime]] 8:24pm up 18 days, 22:43, 4 users, load average: 1.11, 1.11, 1.04 -- David Nicol 816.235.1187 [EMAIL PROTECTED] "I must report that I am in the fortunate position of having logged and categorized my nightmares over the past 37 years." -- Bob Dehnhardt
Re: Backing up IMAP Maildir's ?
I see the question as, "How do I freeze IMAP so it doesn't change anything?" That's usually the backup issue, how to get the file system to hold still while you back it up. OSF1 advfs has a "clone" operation for this purpose, I do not know if other file systems offer similar functionality, of declaring an instant freeze and then tracking changed pages so that life can go on and the backup taken at the moment of the cloning. Without such a facility, you need to shut off whatever causes changes when the backup happens, or resign yourself to failing the consistency-check pass of your backup method. Shutting off mail delivery and imap access (by killing qmail-send and doing whatever it takes to shut off imap service) during a scheduled, planned, outage-for-backups time is one way to do it With plenty of disk space, another possible solution would be to use tar to take a momentary snapshot, such as it is, of your user's situations, and then back up all the username.tar files, which will not be dynamicly changing. That way you don't actually back up (with consistency check) your dynamic user spaces, you back up (to tape) the copies of them. Dave Sill wrote: "Dennis" [EMAIL PROTECTED] wrote: Yes, I'm asking the question again... Why? If you didn't like the first answers, you should say why. Is there a formal way of backing up IMAP Maildir's ? There's nothing magic about maildirs. Your normal backup utilities (tar, dump, etc.) will handle them prefectly well. -Dave -- David Nicol 816.235.1187 [EMAIL PROTECTED] "I must report that I am in the fortunate position of having logged and categorized my nightmares over the past 37 years." -- Bob Dehnhardt
Re: return receipts
Gerry Boudreaux wrote: This is a MUA, not a MTA issue... If the MUA honors the receipt request then the MTA will carry it. Hope this helps Gerry What about the "notification on delivery" stuff -- is that not an MTA feature? Is it deprecated? Rather it would be a feature of the MDA, has anyone added it to qmail-local? -- David Nicol 816.235.1187 [EMAIL PROTECTED] "I must report that I am in the fortunate position of having logged and categorized my nightmares over the past 37 years." -- Bob Dehnhardt
Re: OT: a real MUA for X? (was qmail list reply-to)
Brett Randall wrote: Under X? Try Gnus. It doesn't just work properly in strange situations, it works properly in normal situations as well! don't you have to learn all the saxophone-esque emacs keyboard things to use it? -- David Nicol 816.235.1187 [EMAIL PROTECTED] I don't watch TV, I have no telephone, and I vote
Re: Hard linking messages between maildirs
Paul Jarc wrote: "Slider" [EMAIL PROTECTED] writes: There is an easier solution! If [EMAIL PROTECTED] wants all mail that goes to him to be copied to another maildir as well as for him to get a copy to go to another maildir. That doesn't cover my situation at all. This has nothing to do with delivery addresses. I just want my user agent to copy individual messages, selected by the user, from one maildir to another. paul The way I read the documentation, one of the points of the maildir system is that it allows distribution by hard linking. -- David Nicol 816.235.1187 [EMAIL PROTECTED] safety first: seat-belt wearers for Nader in 2000
Re: having two different routes for pop3 connections
NERvOus wrote: Dear qmail gurus, I have a pop3 server which has got 2 ip addresses and is connected through 2 carriers. Is there a way to let them choose a unique hostname and automagically have them to use mail1.example.com when they connect through isp "X" and mail2.example.com when they use some other isp? allocate a hostname with both addresses as A records. If their pop3 client isn't doing some checking to select the best of the alternatives (I'm not aware of any that do, but it wouldn't be hard to add to an open source one) then they're stuck connecting to the one listed first. I don't know if djdns can be configured to give out A records in a different order depending on the source of the query, if it _can_ and your users are using their ISP-provided DNS servers (they probably are) then you can do it. But altogether this is a DNS issue and belongs on the djdns list which surely exists.
Re: masquerading internal adress for external mail
Since Dave Sill himself didn't come up with the answer to your question, it looks like you're going to have to patch something to do that rewrite for you. What to patch? How to patch it? These are your questions now. Soon, you too will be reading the qmail mailing list, accumulating for weeks in a folder all its own, until you see someone in need of your patch, someone in the same situation you are now in. And you will be able to send them an e-mail saying "Yes, that is possible, in fact I solved that five years ago. Here's a link to my patch!" Davide Giunchi wrote: Hi all. I have a qmail smtp-pop3 server under linux that administer 30 internal user account, this is only an internal server so the internal adresses aren't really present on the net, i would like that if an internal users send an e-mail to the external word the "From:" field would masquerade his internal adress with an unique adress (the only adress that is present on the internet). A friend of mine say me that in sendmail there's this possibility and i'm sure that this is possible in qmail too, how can i do? Thanks Davide. -- David Nicol 816.235.1187 [EMAIL PROTECTED] safety first: seat-belt wearers for Nader in 2000
Re: Open letter
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
"Ihnen, David" wrote:
Maybe an extra-low-effort system would consist of a simply speaking a
keyword into a microphone
I would find this more troublesome than typing my passphrase.
- --
David Nicol 816.235.1187 [EMAIL PROTECTED]
Originator of the world's first combination bassinet and table saw
-BEGIN PGP SIGNATURE-
Version: perl -pe '$_=unpack("u*",$_);'
Comment: 92G5S="!!;F]T:5R(%!EFP@2%C:V5R"@``
iD8DBQE5kDOHJiOJhroV3bkRAtpcAJ4zQtG9qz925plFbbrtWEwveK38LwCeKjnf
/TkbHsLEy4a1ZK+yQ4mYl1k=
=DSp0
-END PGP SIGNATURE-
Re: send to group functionality ?
Is there an easy way to just send to all users in a linux group, instead of having to use an alias file? You mean group, as in, a line in /etc/group, right? Lets say I have a message in a file called syl2000fall.txt and I want to send it to everyone in the group called chem507. I think this would do it: mail `grep ^chem507 /etc/group | cut -f4 -d:` syl2000fall.txt Some light tweaking may be required, but mail will accept comma-delimited recipients, YMMV. -- David Nicol 816.235.1187 [EMAIL PROTECTED] Originator of the world's first combination bassinet and table saw
Re: Sort maildir and send smallest first
I recall from an earlier discussion of a similar problem that a perceived consensus was reached that a good way to do this kind of thing is to patch qmail-remote so that if the message is too big to send, it appears as a temporary error without even attempting to connect. also the definition of "too big to send" is read in from the file system in such a way that it can be easily monitored, and easily manipulated by a cron job. Something like the (non)existence of /var/mail/control/okay-to-send-big. After the status changes, an ALRM signal is generated to reprocess all backed up mail (most of which is there due to being oversize, it is hoped) and that is that. I do not know if this approach can be applied to maildirsmtp and serialmail. Peter van Dijk wrote: On Wed, Jul 26, 2000 at 08:45:34AM +0600, [EMAIL PROTECTED] wrote: [snip] I would like to have qmail changed to do a sort mailbox by seize and send the smallest first. ... The larger messages could then go at night where the x minutes is set to a higher value. -- David Nicol 816.235.1187 [EMAIL PROTECTED] Originator of the world's first combination bassinet and table saw
Re: stop postmaster to make more acounts..
Dave Sill wrote: Geir Ove =?iso-8859-1?Q?=D8ksnes?= [EMAIL PROTECTED] wrote: My customers have paid for like 100 email accounts and one postmaster account... how to i restrict him from making more than 100 email accounts?... this is on a virtual domain.. Run a cron job periodically that removes/disables any .qmail*-default files and any .qmail* files in excess of 100. -Dave better to monitor his overuse and bill him for it -- David Nicol 816.235.1187 [EMAIL PROTECTED] Originator of the world's first combination bassinet and table saw
news: redhat switches to postfix + mailman
This just came in; submitted for your discussion: Attention all List Members: We are in the process of migrating all Red Hat lists from the current list manager software/server running qmail/Smartlist to a new server running postfix and GNU Mailman (*). Within the next week you will receive more information about this migration including information on how to access your membership, including subscribe/unsubscribe information and instructions on how to set your preferences for your membership on the list. The preferences include setting/unsetting digest mode, temporarily disabling delivery, and the option of receiving your own posts. This migration will not only address performace issues that have arisen due to the number of lists/members on Red Hat lists, but also highly improve ease of use for list members. Your patience is greatly appreciated. Thank you for your attention Kambiz Aghaiepour (*) For information on Mailman, see http://www.list.org/ -- \o__O o Kambiz Aghaiepour, RHCE -Phone: (919) 524-7423 o o \_ /|\ -= Red Hat, Inc. =- |\| Pager: (800) 946-4646 //\ //\ |\ |\ -=-=-=-=-=-=-=-=-=-=-=- | | Pager Pin #: 1412622 // // / / |/ mailto:[EMAIL PROTECTED] | http://www.redhat.com |\ ||
shorter file names in the queue
It is true that qmail doesn't do anything with the inode-filename mapping after it is made, besides have unique file names, and that replacing that algorithm (which certainly succeeds in providing insight into How The File System Works) with a different algorithm that also guarantees uniqueness would break nothing? Original Message Date: Thu, 13 Jul 2000 21:38:42 -0700 (PDT) From: dean gaudet [EMAIL PROTECTED] cc:"[EMAIL PROTECTED]" [EMAIL PROTECTED] Subject: (reiserfs) Re: Jedi's qmail reiserfs integration status report On Fri, 14 Jul 2000, Hans Reiser wrote: I strongly believe that it is not deep programming to find a portable version of the algorithm. yeah. i'm not sure how few bits you guys want... but there's the code i put into apache 1.3 for mod_unique_id which generates a 112-bit unique id subject to a bunch of completely reasonable constraints. see http://www.apache.org/docs/mod/mod_unique_id.html for documentation. feel free to snarf the code under any license -- it's essentially unchanged since when i put it in there. it's basically a timestamp, local_ip_address, pid, counter tuple. we replaced qmail's filename generation with this at cp -- mostly because we're still living with solaris 2.6 and it doesn't cache filenames longer than 31 characters; the shorter filenames helped a bunch. -dean
Re: List all users
You mean something like this?
cut -f6 -d: /etc/passwd | xargs -i echo grep "''" "{}""/.qmail-*" | sh
Ari Arantes Filho wrote:
Hi,
Is there some script to list all users including the content of each
qmail-user?
Thanks,
Ari
--
David Nicol 816.235.1187 [EMAIL PROTECTED]
Visualize creamed corn
Re: system rebooted, sendmail took over, how do I send those messages now
Susan Short wrote: Next question, is there another way I can send this mail without getting sendmail to work? One way to proceed is to divide up the queue into messages, and then feed each message into qmail-inject. You will need to "crack" your sendmail's method of storing its queue, but it shouldn't be difficult. Once you have figured it out, feed each message into qmail-inject. -- David Nicol 816.235.1187 [EMAIL PROTECTED] Visualize creamed corn
Recipe For A Good Book On Qmail
Here I go stating the obvious again, but it seems what we are saying is that the archives of this list, plus the documentation that already is in existence, makes up a "book" for those who can access it, ergo what is required might not be an _author_ to create a new, comprehensive restatement but rather an __editor__ to select from the documentation which exists (man pages, LWQ, FAQ lists, et cetera) and compile a compendium. This editor would preferably _NOT_ be someone who "knows a lot about qmail" they would be a quality technical writer with perhaps gardening background. This editor would be assigned the task, they would have to start from zero (zero being a PC with the debian potato installed on it by someone else) and configure and maintain, telling their story and including the documents they find most helpful (with permission of the various authors). That's the first half, in which Linda Potter goes from zero to MTA administrator. For the second half, Linda Potter (our fictionalized hero) builds on the experiences in the first half, going on to install custom patches, to do what exactly? Maybe the first half is all that is needed. Potato to qmail+EZMLM in five or six short chapters. I addition to people participating in this thread on the qmail list, I am CCing a retired technical writer who teaches at UMKC in case this project might appeal to them for collaboration later this summer. Russ Allbery wrote: Rodney Edwards [EMAIL PROTECTED] writes: Qmail ezmlm is now getting so popular that someone has to get their arse in gear and get a book to print. The Idea is a certain winner so com'on O'reilly, Que, or Sam's if your listening in get your finger out guys where drowning out here. I don't believe that publisher interest is the hold-up. To publish a book, someone has to write it first, and one would hope that the people doing so would actually know a decent amount about qmail. :) Those people are somewhat rare; qmail hasn't been around for that long yet. -- Russ Allbery ([EMAIL PROTECTED]) http://www.eyrie.org/~eagle/ -- David Nicol 816.235.1187 [EMAIL PROTECTED] find / -name core|perl -ne'chomp;unlink'
Re: Recipe For A Good Book On Qmail
"John R. Levine" wrote: This editor would preferably _NOT_ be someone who "knows a lot about qmail" they would be a quality technical writer with perhaps gardening background. Having written quite a lot of technical books, I can say that's not likely to work, especially with an editor who doesn't know the material. The qmail list is a swell way to get specific questions answered, but it's a lousy way to get introduced to qmail, and also not a very good way to get answers to questions that aren't phrased concretely. I left out that the "book" would have to go through a beta period of sorts, with public comment, and a "find the bugs" program with money for the first to discover an inaccuracy or make an incremental improvement. It could be The First Bazaar-Style Technical Manual. -- David Nicol 816.235.1187 [EMAIL PROTECTED] JAPH
Re: Qmail on a linux cluster
Does NetFiler let you run other programs on it? I would put one instance of qmail on the netfiler, and insert little tcp-server-protected relay pipes on the other machines to answer port 25. Use maildir. "Matthew S. Crocker" wrote: Hello, I'm building a new mail server/pop server cluster. The cluster will be built using 5 VA linux boxes and a Network Appliance NetFiler. One of the boxes will be a LinuxDirector running the Linux Virtual Server kernel patch. I want all machines to handle SMTP/POP3/IMAP using qmail. User directories will be NFS mounted from the netfiler. Mail queue's can either be local or NFS mounted. David Nicol 816.235.1187 [EMAIL PROTECTED] drawn to the speed and performance
Re: Purpose of this list
Brad Johnson wrote: The other section that doesn't exist (or does it? It's not easy to find) is "Qmail for users" which would talk about qmail just from the perspective of the *nix user, with the userland commands, without mixing it all in with the admin info. there's man dot-qmail -- David Nicol 816.235.1187 [EMAIL PROTECTED] drawn to the speed and performance
Re: distributed redundnat queue architecture (for M Bowman)
Peter van Dijk wrote: this [hypothetical] architecture could result in something similar to usenet Do note that usenet was never designed to guarantee message delivery. Usenet was designed for non-reliable wide-scale messaging. I just meant, it is another architecture where you have redunant data on widely spaced peers. Later, imagining that Bowman wants to guarantee to the corporal in the battlefield who has e-mailed his situation in to HQ, that his smtp 250 from the server means the message will get through even if his immediate mail server gets hit by a missle immediately after sending the code. To implement this, the server would, before issuing the 250, open a channel to a peer and copy the message as far away as possible -- or even attempt immediate delivery -- all BEFORE giving the 250 response. So, 250 on this smtpd does not merely mean "I have taken responsibility for transferring this message" but "I have transferred the message off site already" destination MUAs would need to discard based on redundant message IDs, et cetera I'll stop now...
Re: BACKUP POP SERVER
Make sure you have round-robin turned on in your DNS, assuming that both POP servers have the same name. If that doesn't work, bother half your users and have them change their settings to point to the second machine. I don't see what is saved by this arrangement, over having all the users connect directly to the machine with the mailboxes: all you gain is complexity and additional possible points of failure. NFS isn't free, those packets need to get read off the disk and written to the LAN just the same as if the MUA connects directly. Jhun Hubac wrote: Hi! Is there a way that I can back-up my pop server? I'm using qmail for my two servers (both have SMTP POP3 service). No problem of having redundant SMTP servers but it seems that the MUA (clients) are polling on only 1 of the two servers. I'm using NIS/NFS to distribute information between the two, so their home directories are on a different LINUX machine and the accounts are based on a NIS master. Is there a work-around for this? -- David Nicol 816.235.1187 [EMAIL PROTECTED] drawn to the speed and performance
Re: Disable telnet to port 110
How about a really short time-out? Automated POP3 clients waste no time typing at the prompt -- Mark could analyze the delay his MUAs have between connection and sending auth commands; and patch pop3d accordingly. Or he could patch pop3 to require (not just accept) encrypted authentications, maybe in addition to the timing thing. Paul Farber wrote: I think the original poster is just 'scared' because the POP3 protocol uses cleartext command (telnet, perl script, python) could connect up and get mail. Thinking that telneting to 110 and giving the same commands at a console is somehow 'hacking' a system. It will blow thier mind when they telnet to port 25 and can actually SEND mail! Paul Farber Farber Technology [EMAIL PROTECTED] Ph 570-628-5303 Fax 570-628-5545 On Mon, 15 May 2000, Aaron L. Meehan wrote: It's difficult to answer a nonsensical question. Aaron At 05/14/2000 05:48 AM Sunday, Mark Lo wrote: I would like to disable telnet to port 110, but still let my user to retrive mail via mail client at port 110?? (using tcpserver) -- David Nicol 816.235.1187 [EMAIL PROTECTED] drawn to the speed and performance
distributed redundnat queue architecture (for M Bowman)
Michael Boman wrote: A server goes down [and the mail should been taken care of by another server, automatically and samlessly.] A single point of failure is not an option. Best regards Michael Boman At the cost of more WAN traffic, you could add patches so that on delivery failures, in addition to a message being added to the local queue it also gets copied to one or more other peers for queuing. Whenever a message that was queued gets successfully delivered, a notification message is sent to the associated peer, so it can dequeue its redundant message. Implementing this would require: full description of the redundancy protocol implementing the protocol in the software Depending on the various costs (WAN bandwitdh, CPU, storage space, programmer time) this architecture could result in something similar to usenet, with each extended queue storage server contacting the others at regular intervals with a list of message-IDs it has received, so that all of them get multiple chances to receive the same message stuck in the queue. -- David Nicol 816.235.1187 [EMAIL PROTECTED] drawn to the speed and performance
Re: spool vs individual files
Dave Sill wrote: Large mbox mailboxes are huge, unwieldy files. Large maildir mailboxes are huge, unwieldy directories--most filesystems don't handle them efficiently after they grow to a few thousand files. DEC/Compaq ADVFS handles huge directories without trouble, using a hashed directory table that grows as needed. For linux, reiserfs supposedly doesn't care about directory size as it blurs the directory_entry / file_data distinction. I should have qmail running over reiserfs any day now, and will give a report when I do -- David Nicol 816.235.1187 [EMAIL PROTECTED] You discover uranium: collect $240,000
Re: origins of Bracketed Quad notation
Peter van Dijk wrote: [the 821] definition is incorrect, in that it allows stuff like [10.10.10.1].vuurwerk.nl I think this was superseded in a later RFC. Thanks, all! I wonder if Postel meant for constructions such as Peter's error to signify numeric addresses internal to private networks. This would be in keeping with 821's emphasis on source routing. __ David Nicol 816.235.1187 [EMAIL PROTECTED]
origins of Bracketed Quad notation
Can anyone point me to the IETF RFC describing e-mail addresses of the form david@[10.10.10.10] Although web pages refer to this construction as a "821-compliant address" I found no discussion of referring to hosts by anything other than names within 821. -- "Lord Macbeth knew he was approaching the SITE of the rout from the SIGHT of odd body parts scattered on the blasted heath."
Re: Antigen found =love-letter-for-you.txt.vbs file
Kai MacTane wrote: At 5/5/2000 11:54 AM -0600, ANTIGEN_HOUSTON wrote or quoted: Antigen for Exchange found LOVE-LETTER-FOR-YOU.TXT.vbs matching =love-letter-for-you.txt.vbs file filter. The file is currently Detected. The message, "Re: hack for filtering "i love you" worm", was sent from Kai MacTane and was discovered in IMC Queues\Inbound located at Matchlogic/MATCHLOGIC/HOUSTON. Hmmm. Looks like someone's already filtering on just the string I sent out. I wonder if they're filtering all .vbs files? Our exchange admin is. __ David Nicol 816.235.1187 [EMAIL PROTECTED] "Lord Macbeth knew he was approaching the SITE of the rout from the SIGHT of odd body parts scattered on the blasted heath."
Re: Future of qmail: will it care about viri/worms/etc?
Keith Warno wrote: there should be no need to "hack" qmail And there isn't! Why do people persist on insecure MUAs? __ David Nicol 816.235.1187 [EMAIL PROTECTED] "Lord Macbeth knew he was approaching the SITE of the rout from the SIGHT of odd body parts scattered on the blasted heath."
multiple rcpt patch idea etc
Dave Kitabjian wrote: Regarding: http://web.infoave.net/~dsill/lwq.html#multi-rcpt Dave S, I'm having trouble accepting this logic. You mention 3 options: "Say you're an MTA, and one of your users sends a message to three people on hostx.example.com. There are several ways you could do this. 1. You could open an SMTP connection to hostx, send a copy of the message to the first user, send a copy to the second user, send a copy to the third user, then close the connection. 2. You could start three processes, each of which opens an SMTP connection to hostx, sends a copy of the message to one of the users, then closes the connection. 3. You could open an SMTP connection to host, send a copy of the message addressed to all three users, then close the connection. " and that qmail uses option #2. Clearly, the rank of efficiency is, from best to worst,: 3, 1, 2 You have analyzed the situation (or part of it at least) correctly. Thing is, Dan optimized for SECURITY not EFFICIENCY. There exists or may exist a class of broken MTA which cannot process multiple receipts correctly; or which leaks bcc information during a multiple receipt. I wrote some code that pre-invokes qmail-remote, feel free to give me credit when you use it, it is at http://www.davidnicol.com/qmail.html and I will be revising is as needed (into its own file most likely) Before I patch qmail-smtpd to do essentially that preprocess when there are multiple recipients, instead of whatever it normally does, somebody talk me out of it? My thought is, if mail only gets into the queue after it has been attempted once, then mail in the queue has already failed at least once and properly should be attempted in trickles. And also Chris Hardie writes: Clearly it's a complicated issue, but it seems that as broadband access to the net becomes more common, businesses are going to expect to be able to use one "interface" to do all their communications, be it plain text messages or large multi-megabyte file transfers. I cringe every time someone sends me a 7 MB mail message, but it's difficult to explain to them why this is a bad idea. I'd be interested to hear if anyone's found a good general solution to this in a production/business environment. One approach might be to rig the MTA to unpack attachments, give them unique and secret file names, store them in per-user directories where a http server on the MTA host can see them (but not server directory indexes) and replace the attachments with links to the files. This would have the effect of server-side-selecting the "view attachments as links" option present in some MUAs. Fine-grained administrative control could be asserted over how much space in e-mail attachments you may have before the last used gets cleared to make space, and so forth. This is what Scott Gifford suggested, although he wanted to add password-protection instead of giving unique, random file names. __ David Nicol 816.235.1187 [EMAIL PROTECTED] "Lord Macbeth knew he was approaching the SITE of the rout from the SIGHT of odd body parts scattered on the blasted heath."
VMS mail.mai files?
Anyone got any .mai conversion tools? __ David Nicol 816.235.1187 [EMAIL PROTECTED] "Lord Macbeth knew he was approaching the SITE of the rout from the SIGHT of odd body parts scattered on the blasted heath."
Re: opinion on my proposed setup!
Madhav wrote: This setup actually provides protection from any failure. Once it's running, pull the plug on the nfs server...
Re: mail-abuse.org
Peter van Dijk wrote: On Mon, Apr 17, 2000 at 07:08:39PM -0400, Len Budney wrote: "Luis Bezerra" [EMAIL PROTECTED] wrote: [edited out] Or are you an autoresponder? If so, are you available under the GPL? Maybe I can run you from procmail, to annoy people who annoy me. If not GPL, then we'll just reverse-engineer him and open the source up anyway :) ROTFL
Re: using perl to send a list message (with qmail)
Not tested. May provoke unpleasent censure. Use at own risk.
open (LIST,"/home/justin/www/cgi-bin/lovely_people.list");
foreach (LIST) {
#send mail to "$_"
}
2)this list has about 1000 people, and it takes more than a couple minutes
for perl to go through that all,making my program look like its hanging
There's always
fork and goto GLADTHATSOVER
close STDOUT;
open (LIST,"/home/justin/www/cgi-bin/lovely_people.list");
foreach (LIST) {
#send mail to "$_"
}
exit;
GLADTHATSOVER:
And since you've now got all the time you need to process the list,
you can even save bandwidth by arranging them by destination host
and invoking qmail-remote directly:
%Recipients=();
while(LIST){
my($name,$host) = m/^\s*([^@]+)\@([^@]+)\s*\Z/ or next;
$Recipients{$host} = [ $name.'@'.$host, @{$Recipients{$host}}];
};
for(keys %Recipents){
$/ = "\0";#match qmail-remote's idea of how to end a report item
open QR,"/var/qmail/bin/qmail-remote $_ justin\@summerspam.com
@{$Recipients{$_}}lovelymessage|";
while(QR){
$This = shift @{$Recipients{$_}};
next if $This =~ m/^K/; # delivery successful
# otherwise, queue it up for later
system"/var/qmail/bin/qmail-inject -r -fjustin\@summerspam.com
$Thislovelymessage|";
};
};
Good lord! I'm instructing the world how to send thousands of e-mails
from within a CGI! Please use this for good and not for evil!
I'd better insert some tactical syntactic problems...
While we're at it, would anyone like some blue prints for atomic
weapons?
delivery hiccup involving MDaemon.v2.7.SP4.R and hacked 250 reply
I'd like to think it's my fault, for altering smtpd.c to have cute messages, but this has never happened before, so it could mean that MDaemon got confused by the non-standard 250 code, saying more than "250 ok" like others do, and returned the buffer in a rcpt to: command. I'm looking at these headers: Received: (qmail 501888 invoked by alias); 10 Apr 2000 20:55:05 - Delivered-To: 250 ok yes [EMAIL PROTECTED] Received: (qmail 501889 invoked from network); 10 Apr 2000 20:54:54 - Received: from asub.arknet.edu (HELO asubserver1.asub.arknet.edu) (150.208.48.1) by tesla.umkc.edu with SMTP; 10 Apr 2000 20:54:54 - Received: from guest.asub.arknet.edu [150.208.48.186] by asubserver1.asub.arknet.edu [127.0.0.1] with SMTP (MDaemon.v2.7.SP4.R) for [EMAIL PROTECTED]; Mon, 10 Apr 2000 15:46:50 -0500 Apparently qmail on tesla delivered this to "250 ok yes [EMAIL PROTECTED]" which fell through to me, instead of queueing it and re-sending it to [EMAIL PROTECTED] like was supposed to happen. I have forwarded the message to the correct recipient, but I would like to prevent this from happening again. I am guessing that I need to include a dash after the OK in the replies, like microsoft exchange has, to fix this. Reviewing the relevant documentation, that is, http://cr.yp.to/smtp/request.html, I am led to understand that a SMTP response of "250 ok yes sir" is every bit as valid as exchange's "250 OK - Recipient [EMAIL PROTECTED]" so I am not changing my server. Thoughts? _ David Nicol 816.235.1187 [EMAIL PROTECTED] "Nobody wants a hammer with racing stripes and a horn." - Greg Knauss
Re: Encryption and t-shirts
Greg Owen wrote: Or how about Front: "Don't queue mail with sendmail" Back: "Send mail with qmail" ROTFL. I'd buy that one. If someone else can do the art I can get them printed ___ David Nicol 816.235.1187 [EMAIL PROTECTED]
Re: How to know how to display a email...
Michael Boman wrote: Problem with the solution: How the heck can I in a email see if I need to display it as english, chinese or japanise text? This question has ZERO to do with the operation of the MTA (beyond if it is passing 8-bit unicodes cleanly or suchlike.) I'd ask comp.lang.perl.misc, assuming you're going to be doing this parsing using Perl. ___ David Nicol 816.235.1187 [EMAIL PROTECTED]
reiserfs
Anyone have any reiserfs stories? ___ David Nicol 816.235.1187 [EMAIL PROTECTED] e is one key to the right from w
Re: Restrict Times
Director tecnico del Nodo Nicarao -- Juan Navas wrote: Hi, I was wondering if any of you know of any qmail feature that allows restrict E-Mail checking at a specific time of the day Juan Navas System Administrator Managua, Nicaragua The general solution to this kind of thing is to hack and add, if not already available, a feature by which the behavior you want to modify is controllable by a configuration file, and then to write two scripts to modify the configuration file for on and off, and invoke the scripts from the crontab. Don't know what you mean by "E-Mail checking" so can't be any more specific. ___ David Nicol 816.235.1187 [EMAIL PROTECTED] I would vote for a Trump/Hightower reform ticket
Re: Corel Linux ships with qmail installed, but not running
"Chris L. Mason" wrote: Perhaps Corel is planning to use qmail in future versions and it just wasn't ready for 1.0? I've been waiting awhile for a Linux distribution to come out that uses qmail as the default MTA (or at least offers the choice of using it over sendmail in the installation.) Chris I think linux distributors are waiting for qmail to be able to combine non-VERP messages for same remote machine into single transmissions; or perhaps for a configuration GUI. (it would thave to be standard and extensible, so qmail extensions could extend the GUI as well) Where's your distribution, then, Chris? Does not the existence of easily available qmail RPMs qualify? __ David Nicol 816.235.1187 [EMAIL PROTECTED] grep -v 0 /proc/*/where
Re: Should qmail immediately reject relaying? [was Re: Qmail is relaying external mail]
what keeps spammers from faking envelope-from and using include-in-bounce features to relay spam content? Is it possible that a subject of "failure notice" will some day not be sufficient to prevent this possibility? I HEREBY PATENT THE METHOD!
Stopping big messages until later
Thinking about this situation over the weekend I concluded that
the sanest thing to do would be to hack qmail-remote so it
checks file-size and marks temporary failures for oversize mails during
peak times. This could be done by reading a file size from an external
file, and having the file size modified by something run from cron; or
the policy could be written into the patch (which would make the patch
too site-specific to share.)
Looking at the the qmail-remote.c program, I suppose the patch would
first define an error handler like all the other error handlers :
void temp_policy_size() { out("Z\
Message is too large to send during peak time. (#4.3.0)\n"); zerodie();
}
Then, find a place where we can stat the message text No, we can't
at all because qmail-remote reads stdin -- although we could read stdin
until we have enough, and issue the error the same place djb issues the
"out of memory" temporary system error --
no, we appear to be reading stdin and passing it to remote in
little pieces? The sending is done within the smtp() call -- in which
we learn that a new temp_* message is a breakable convention, as the
cquit("Z..") will work too -- or is that just within the smtp
session?
Anyway, the message gets sent within blast(), which reads from ssin,
doubles line-starting dots, and writes to smtpto, one character at a
time.
So.
In order to implement a size-driven policy completely within
qmail-remote,
you could allocate a buffer large enough to hold an entire
size-compliant message and load it from stdin, before looking up hosts
and dns and so forth. If you get to the end of my buffer and we are
doing
size-rationing, I would temp_policy_size() and that's that.
Later, when the smtp session is under way, blast() would read and send
the
buffer, before reading stdin, if there is any stdin left to read,
because
we're not in peak hours.
Alternately, you could have two qmail-remote programs, the original one
and
one that reads into a limited buffer, and then refers to it within
blast()
instead of reading ssin, and have your cron job switch the link at
/var/qmail/bin/qmail-remote between /var/qmail/bin/qmail-remote_original
and /var/qmail/bin/qmail-remote_size-policy at the transition times,
something
like
0 8 * * * ln -f /var/qmail/bin/qmail-remote_size-policy
/var/qmail/bin/qmail-remote
0 20 * * * ln -f /var/qmail/bin/qmail-remote_original
/var/qmail/bin/qmail-remote
Since qmail-rspawn finds the program to run by having execvp consult the
file
system, this will work. To be supersuper safe you could add a wait and
retry
once before exiting with error condition somewhere near the execvp in
qmail-rspawn.
That way you don't have to mess with new control files any.
There are some rough edges in there; when I write my qmail-inspired MTA
I think I'll have the qmail-remote analogue take a file name as an
argument
and be responsible for setting the delivered flags itself instead of
simply reporting delivery status; this will of course make my system
vulnerable to stack-smashing attacks from remote SMTP servers, which
would
then have sufficient privilege to do some damage, should the OS I will
be
using be vulnerable to such things.
Ari Arantes Filho wrote:
David,
You are totally right:
What ari appears to me to be asking for is a way to derail large e-mails
into a secondary queue: He wants email to flow 24/z for little memos,
but attachments above a threshold must wait until off-peak.
I'm using the qmail-hold patch, so I can create a control/holdremote (with
1), send HUP the qmail-send and the queue is paused. But at this time, all
messages will be stopped. I would like to stop only the big one.
Re: Stopping big messages until later
[EMAIL PROTECTED] wrote, but did not CC everyone:
On Tue, 14 Dec 1999 01:57:59 + , "David L. Nicol" writes:
Looking at the the qmail-remote.c program, I suppose the patch would
first define an error handler like all the other error handlers :
void temp_policy_size() { out("Z\
Message is too large to send during peak time. (#4.3.0)\n"); zerodie();
}
Then, find a place where we can stat the message text No, we can't
at all because qmail-remote reads stdin -- although we could read stdin
until we have enough, and issue the error the same place djb issues the
"out of memory" temporary system error --
Actually, you could just use fstat() on stdin. That
would give you the file size, since qmail-remote's
stdin should be a regular file.
I like the idea of modifying external state from a
cron job -- that makes it much more flexible.
The one gotcha is that qmail-send's back-off might
skip the "big message window," so you should probably
ALRM qmail-send when you switch policy.
--
Chris Mikkelson | Setting delivery schedules is easy enough using the
[EMAIL PROTECTED] | I Ching, astrology, psychic hotlines, or any of the
| well-known scatomantic and necromantic methodologies.
| Meeting your prophetic deadlines, though, is another
| bowl of entrails. -- Stan Kelly-Bootle
Re: How to send a message after
What ari appears to me to be asking for is a way to derail large e-mails into a secondary queue: He wants email to flow 24/z for little memos, but attachments above a threshold must wait until off-peak. A variety of approaches come to mind. Disabling _all_ outgoing e-mail until off-peak times is not one of them. Some point in the process must be selected for removal/insertion, and a size-based gateway installed there. Read your source code, Ari, I look forward to seeing your upcoming patch! Ari Arantes Filho wrote: The user sends the email, not me!!! - Original Message - From: [EMAIL PROTECTED] To: Ari Arantes Filho [EMAIL PROTECTED] Cc: [EMAIL PROTECTED] Sent: Friday, December 10, 1999 1:42 AM Subject: Re: How to send a message after run your mail program in night;-) On Thu, 9 Dec 1999, Ari Arantes Filho wrote: Hi, Supose I'm supervising the qmail queue and see a message with (2 attachments of 2mb each for 5 different addresses). This message will consume a lot of my link, so I want to send this message during the night. How to do this? Best regards, Ari -- ___ David Nicol 816.235.1187 [EMAIL PROTECTED] Safeguard your finances against the great Beanie crash of 2000
Re: secondary mail relay: rcpthosts AND SMTPROUTES
Steve Vertigan wrote: if it was really a lower priority why did the error message begin "I am listed as the *primary mx* for this host"? Because there's a bug in the way the determination of "primary MX" is made. I have not looked at the source code of how the determination is made. $ dig mx umkc.edu ; DiG 2.2 mx umkc.edu ... ;; ANSWERS: umkc.edu. 86400 MX 5 email.exchange.umkc.edu. umkc.edu. 86400 MX 20 umx.missouri.edu. umkc.edu. 86400 MX 200 134.193.4.60. I suppose that now someone is going to try and tell me that 200 5 ? __ David Nicol 816.235.1187 [EMAIL PROTECTED]
Re: Sendmail Virtusertable equivalent?
Peter Gradwell wrote: [1] Which is why, if your mail server is the best MX preference host How does qmail make this determination? Does it get the preference fields from the dns and choose the lowest one, or does it rely on a system call? qmail-remote.c refers to subroutined defined in dns.c, but the comments are sparse
Re: secondary mail relay: rcpthosts AND SMTPROUTES
"Timothy L. Mayo" wrote: domain.tld. 86400 MX 200 nnn.nn.nn.nnn ^ This is your problem. An MX record may ONLY point to a A record machine name. Fix your DNS and I can guarantee that the rcpthosts-only entry will work. Hmm. having readjusted the dns to serve a name instead of a number on a test domain, it does not appear to bounce. I am not removing my smtproutes entries, to reduce dns load, and to prevent messages getting forwarded around between multiple secondaries. I would like to see qmail-remote.c adjusted to account for this particular flavor of misconfiguration, which is clear enough to enough MTAs to cause the machine at nnn.nn.nn.nnn to receive plenty of e-mails, yet which causes qmail-remote.c to mistakenly determine that the local machine is the best-choice server for a domain, in release 2.0.
Re: secondary mail relay: rcpthosts AND SMTPROUTES
"Timothy L. Mayo" wrote: On Mon, 6 Dec 1999, David L. Nicol wrote: And add a line in control/smtproutes too; otherwise you'll bounce messages as qmail mistakenly interprets that it is supposed to be the end recipient. This starts happening only after you actually modify the MX records. No. An smtproutes entry is NOT needed. The only time you would have a problem would be if you placed your server at the same MX or higher priority as the machine you were serving as the secondary for. (Remeber that a HIGHER MX number is a LOWER priority.) - Timothy L. Mayo mailto:[EMAIL PROTECTED] Senior Systems Administrator localconnect(sm) http://www.localconnect.net/ Yes, that is what I thought, too, until I did it. The primary MX has priority five, the secondary has priority 20, and I set the qmail box to have priority 200 and what happened to the occasional piece of e-mail that got to it? It was bounced, with a message that said "Although I am listed as the primary mx for this host, I haven't a clue what to do with this piece of e-mail." (from memory.) After concernedly rereading the FAQ I added lines to smtproutes and things are now working properly: the occasional piece of overflow that wanders into the box in question is now held briefly and then forwarded. The fact that I had no "locals" file may have had something to do with it; although the documentation seems to say that a locals file is not needed if you only accept local mail for "me." The moral of the story? Set up test cases before altering your production systems, no matter how well-documented and "authoritatively" asserted the feature may be. __ David Nicol 816.235.1187 [EMAIL PROTECTED] Corel Linux is Debian with qmail preinstalled
Re: Rewriting without date header in a perl oneliner
I read the documentation of qmail-inject and it would do what I want if I could filter out the old "Date" line. I can't (almost) write in Perl. I can write a C program. But first I wanted to know if there is some already done script/program which would delete a chosen line from RFC822 header. I don't want to reinvent a wheel. I've been reinventing wheels for decades. Here's a perl one-liner for you to strip out the first date header and pass everything else: (will be odd but not fatal if the date header wraps.) perl -ne 'print if($m or ($m = /^Date:/ and next) or ($m = ! /\w/) or 1 )' Which goes through every line in the file and prints it, except for the first time the line starts Date, unless of course if that is after the first blank line. __ David Nicol 816.235.1187 [EMAIL PROTECTED] End Daylight Savings Time in the year 2000 --just say NO
Re: qmail remote delivery logic
We have the source; let's fix it. What the people with the problem are asking for appears to be for qmail to not split up identical mails intended for multiple recipients at identical hosts. These are real problems and poo-pooing them as degenerate cases or something produces nothing. In terms of modifying, this might not be the "extensive rewrite" that "life with qmail" claims it will be. I see two parts to change: We want (1)the part that splits messages with multiple recipients to group by mail-host-name and merely split by mail-host-name, and also (2) that qmail-remote can issue multiple rcpt-to instructions in these cases. That is all. Two patches. Three, with (3) record-keeping regarding who has received and who has errorred adjusted to work with (2.) People who are running mailing lists (which need VERPS) behind low-bandwidth links are not covered by this patch proporal: They need to form cooperatives and rent servers with good connections. Looking at the chart qmail-smtpd --- qmail-queue --- qmail-send --- qmail-rspawn --- qmail-remote / | \ qmail-inject _/ qmail-clean \_ qmail-lspawn --- qmail-local it is not exactly clear at which point a mail that is CCd to [EMAIL PROTECTED], [EMAIL PROTECTED], [EMAIL PROTECTED] gets split into three messages. (by smtpd, queue or send?) man pages indicate ... that qmail-remote "sends the message to one or more recipients at a remote host." Which means that it still hasn't been split up when qmail-remote gets it, and that qmail-remote is the only program that would need to be patched. Is this accurate, that messages withmultiple recipients are associated with a single queue entry until they are delivered and cleaned up, and that all delivery multiplexing happens within qmail-remote? If so, qmail-remote is the only part of the system which needs to be tweaked, and the groundwork is already there. Pavel Kankovsky wrote: On Mon, 8 Nov 1999, Dave Sill wrote: I think Postfix just sorts by FQDN, so it doesn't have to do 10,000 DNS lookups before it starts delivering. But by doing that, it potentially misses a lot of combining for different FQDN's with the same MX. "A lot" being a speculation or based on real-world data? evil grin qmail-remote has to look up the MX records too, so adding a switch to just sort by host name or wait until all the mx queries are back before sorting would not be that hard; but this advanced optimization would only make sense as something to toss in AFTER sorting and grouping by hostname is in place, at that point it's simply adjusting the sort/group method, it isn't introducing any new architectural features. __ David Nicol 816.235.1187 [EMAIL PROTECTED] End Daylight Savings Time in our lifetime
qmail - PMDF weirdness
The situation is, that messages for certain (but not all) recipients on a VMS-PMDF system do not get delivered from qmail. PMDF issues odd error messages or drops the connections, on only these users. Messages to other users go through fine. I have a sloppy working fix of setting up a smtproute to the PMDF box through a relay running (boo, hiss) sendmail. Messages to these same users neither bounce nor hang up in sendmail's queue: these users don't have full mailboxes or something like that. Does anyone have an idea what might be happening and how to fix it? Here are qmail's two styles of log entries from the failed attempts: Oct 26 00:46:51 gungadin qmail: 940916811.156160 delivery 30: deferral: Connected_to_134.193.4.1_but_ connection_died._(#4.4.2)/ Oct 26 14:45:45 gungadin qmail: 940967145.640832 delivery 60: deferral: 134.193.4.2_failed_after_I_se nt_the_message./Remote_host_said:_421_4.4.2_Timeout_while_waiting_for_command./ The PMDF site has a mention of qmail relating to the periods which end SMTP data, but that is all. The messages involved in these failed delivery attempts were identical to messages to other PMDF users which went through without difficulty. __ David Nicol 816.235.1187 [EMAIL PROTECTED] "ereiamjh"
X-Face headers
Russell Nelson (apparently a xfmail user) wrote:
X-Face:
$K'YURj"g6ImvqTS_=]8)gqh!5;ElY[.Rao%j8r+]iUfE{%|v%F=mcq6l{K=~mf#:?"
nslS]U~|x{2V=Eex_I#"9K~9)?m7Lm={(j_)SX~fzgST~P%QUhc{1p]c3@Zn1u*PZlkHM**X^vV
lGkB5y^Kz%w5p~^uDue]hLke,N;+QImMCdCr~Kz--?|SS?DbZiaE;xPW/7k9u_cc(It%mvMNVk;
qVk~
Is there any way to see these if you don't have XFmail? Would
a preprocessor that converts them into attached inline MIME gifs
be too too too tricky? Procmail rule? Pestering Mozilla.org
wish list to replace the Customizable N-Thing with the X-Face
would result in much greater popularity of X-Faces, re-tacking
them onto messages as a trailing (or leading) inline picture would
work too and wouldn't require changes to MUAs that already can
see inline graphics.
__
David Nicol 816.235.1187 [EMAIL PROTECTED]
It is difficult to tell if your employees are doing real work
or just goofing off when tools and games have the same GUI.
-- Dennis Chao
RFC: backoff enhancement idea
I have one user whose mailbox drops connections on a machine that is up and generally accepting mail. Currently this user has five messages waiting to be sent to them, each one getting progressively longer retry times. I would like to see the retry time for new message for a remote address that already has a temporary failure associated with it start at the current longer delay time, instead of each message backing off on its own schedule. Sound good? Would this require much additional space within qmail-send? Perhaps a file of backoff times could be maintained and read only when a delivery has temporarily failed, to prevent additional memory use by qmail-send What thinks everyone? ___ David Nicol 816.235.1187 [EMAIL PROTECTED] Enough already.
Re: Any thoughts on instant messaging vs. smtp
Eric Dahnke wrote: I understand the pros and cons of each, but am interested in knowing if there is anyone on this list who thinks instant messaging has a chance of upseating smtp. - cheers Eric "talk" is as least as old as SMTP. Did the appearance of the telephone eliminate the postal service? Hardly. ___ David Nicol 816.235.1187 [EMAIL PROTECTED] Enough already, Noam Chomsky for president
binmail on tru64 version 5.0
bin/mail on osf5 / DU5.0 / Compaq trucluster64. tru64 binmail cannot take the -f switch, as -f means something different here than it means to the system 7 syntax as given in all the /var/qmail/boot examples. I have determined that I can get a successful qmail delivery using OS-proved /bin/mail with this /var/qmail/rc file: #!/bin/sh # Using splogger to send the log through syslog. # Using binmail to deliver messages to /var/spool/mail/$USER by default. # on tru64 bin/mail, -f is a switch to specify an alternate mbox file # so use -d only exec env - PATH="/var/qmail/bin:$PATH" \ qmail-start \ '|preline -f /bin/mail -d "$USER"' \ splogger qmail ___ David Nicol 816.235.1187 [EMAIL PROTECTED] Enough already, Noam Chomsky for president
Re: spambait?
I'd like to play too -- is there an address to forward my trapped spam to, if I was to set up a couple clearly marked spam addresses and stick them on my web pages? I have control over virtmaps and aliases files on several domains right now "John R. Levine" wrote: The closest automated thing is the MAPS RSS which lists open relays that send spam. Many spam traps (including mine) autoforward stuff for testing and listing. To prevent spoofing, people who the manager knows get passwords to put in the submissions that let them bypass his manual scrutiny. It works pretty well, blocks a lot of spam for me. __ David Nicol 816.235.1187 [EMAIL PROTECTED] End Daylight Savings Time in our lifetime
Re: Command-line mailer
mpack can be incorporated into your packing schemes, instead of using one part of a larger higher=level abstraction http://filewatcher.org/sec/mpack/ Jason Haar wrote: Sending an attachment as you do requires something more sophisticated than mailsubj. I'd use mutt - darn near the best mailer money can't buy!!! http://www.mutt.org/ __ David Nicol 816.235.1187 [EMAIL PROTECTED] End Daylight Savings Time in our lifetime
Re: extracting passwords from NTMail?
Theodore Cekan wrote: I will be converting our NTMail installation to qmail. Does anyone know if there is a way to extract passwords in plain text from NTMail? Thanks, Ted Keep the NTmail server up during a transitional period, and write a fallback script to your password database that checks mismatches against the NTmail server using pop3. You can even get the libntlm package and use NTLM authentications during your transitional period but that is may be silly. So you aren't extracting them, but trapping them, which is more work but means the users jsut have to log in during the transitional period instead of assigning them new passwords at transition time. POP3 checkers are easy to write, I can send you UMKCs if you need. libntlm is by Grant Edwards; is used in recent Fetchmail releases; his FTP space is ftp://ftp.visi.com/users/grante/stuff __ David Nicol 816.235.1187 [EMAIL PROTECTED] End Daylight Savings Time in our lifetime
