Re: What about www.mail-abuse.org ?

2001-06-03 Thread Mark Delany

On Sun, Jun 03, 2001 at 09:04:22PM -0700, Tupshin Harper allegedly wrote:
 My test of your server indicates that you appropriately block relaying.

(Let me say beforehand that I don't know anything about mail-abuse.org
and whether they do or do not have this address listed, or indeed
whether they have this address listed for valid reasons).

The fact that some IPs are not accepted for relaying does not mean
that all are. It may well be, for example, that the IP in question
relays mail from, say, all 202. addresses or all 202.96 addresses.

Of course this is not a qmail related issue unless the original poster
has a problem understanding relay protection with qmail and starts
with a posting of his tcpserver rules and his expectations of what
they do.


Regards.



 
 -Tupshin
 
 - Original Message -
 From: daiyuwen [EMAIL PROTECTED]
 To: [EMAIL PROTECTED]
 Sent: Sunday, June 03, 2001 9:11 PM
 Subject: What about www.mail-abuse.org ?
 
 
  Hi, Dear All
 
  Somebody are talking about www.orbs.org.
  What about www.mail-abuse.org?  I think they're abusing their influence.
 Many sites are using their blacklist.  So they should be very responsible
 for every IP address their list.
 
  For my instance, my server is on the RSS list because it WAS an open-relay
 server.  Then I fixed the problem and sent a removal request.  But
 mail-abuse.org said I blocked their mail server (I didn't.  I don't know
 why).  Now they even refuse my removal request on the web.  According their
 order, I had to mail to [EMAIL PROTECTED], explaining why I blocked their
 server. But I just got an auto-relay that said I should submit removal
 request on the web.  Dead loop :-(
 
  Any body kind enough to test if my server is third-party relay?  Its IP
 address is
  202.96.230.197
 
  Best regards,
  Dai Yuwen
  __
 
  ===
  ÐÂÀËÃâ·Ñµç×ÓÓÊÏä (http://mail.sina.com.cn)
  ʹÓÃÊÖ»ú¶ÌÐÅ¡°ÓʼþÌáÐÑ¡±¹¦ÄÜ£¬ËæʱÁ˽âµÄÊÕÐÂÐÅÇé¿ö£¡
 (http://sms.sina.com.cn/docs/sina_mailalert.html)
  ¶©ÔÄÊÖ»ú¶ÌÐŶ¥¼¶ÐÂÎÅÿÌìµÃпîÊÖ»ú´ó½±£¡
 (http://dailynews.sina.com.cn/c/266499.html)
 
 



Re: qmail on SCO OpenServer

2001-06-03 Thread Mark Delany

On Mon, Jun 04, 2001 at 02:16:20PM +1000, Jason Heskett allegedly wrote:
 Hi there,
  
 I am probably opening a long-running topic here, but here goes...
 I have just successfully compiled qmail on SCO OpenServer. However, it seems
 that my outgoing mail queue is getting stuck.

Is that true for all outgoing mail or just some?

 The log includes, 
 Connected_to_..._but_connection_died._(#4.4.2)/
 
 Running a ps shows qmail-remote sitting there, trying to deliver the
 queue.

Does SCO has a truss or strace or some similar system call trace? If
so, attach to the qmail-remote and show us the output. Yo may also
want to get a tcpdump/snoop of the tcp traffic.


 Local deliveries work just fine.
  
 I know similar messages have been posted to the list, and I apologise for
 the duplication,

You'll also note that SCO in general is not well loved/supported by
djbware. The problem seems to be that the tcp/ip stack sucks - to use
a technical term.

 Before you say anything I can't move to Linux just yet...

That still leaves any of the BSD variants then :


Regards.



Re: do I need to log

2001-06-03 Thread Mark Delany

On Tue, Jun 05, 2001 at 02:28:31AM +, NewBiePortal allegedly wrote:
 
 Hi 
 
 I'm wondering, do I really need to log anything. Is this must or is it extra for 
debugging purpose. I just feel that there would be much improvement with the sending 
mail if my cpu did not have to bother with logging every email that's leaving my 
mailer.  I mean I have millions of junk emails which none of them are important at 
all.
 
 I'm kinda of newbie but can someone confirm that It's okay to get rip of 
 qmail-smtpd/log/run

It's entirely up to you. I wish I was lucky enough to work on an email
system that has millions of junk emails and which required no
analysis or problem diagnosis or anything, ever! Just remember most
problems have to be looked back at which is only possible with some
sort of log.

Of course the fact that your system does have millions of junk
emails suggests that something is very wrong in the first instant -
something like being abused as an open-relay that a log might well
identify.. But as I say, it's your server.

There is the final point that you don't know what your logging really
costs. How much of a bother is it to your CPU? Have you measured it
or are you speculating? Is the bother greater than that or the
millions of junk emails that you might be able to eliminate?


Regards.



Re: Oops,I guess Sendmail wasn't secure after all...

2001-06-02 Thread Mark Delany

On Sat, Jun 02, 2001 at 05:20:01PM +0200, Boris allegedly wrote:
 Hello Johan,
 
 
 JA Not quite. More like someone inspects your free car and finds a button
 JA that can make it explode. Maybe he pushes the button, maybe not. Maybe he
 JA pushes the button on someone else's car. Are you willing to take that
 JA risk? I can imagine two situations where that would be the case: either
 
 Well, there is no button with a text like press me here -) for
 the public.

Of course there is, silly.

Tell us, your mail progam seems to be The Bat! (v1.48f) Personal -
did you write this program from scratch yourself or did you simply
click a few buttons and install the work of someone else?

Now, what do you think most script kiddies do? They don't scour the
code for exploits as you imply with there is no button. They simply
download the hard work of one or two people and install the pre-built
button. It's trivial. So, press me here is as far away as a
download. You're not seriously suggesting this is a serious secruity
barrier are you?

 If we are talking about the security of a product, we have several
 things to take a look at. Internal security (a mailserver-only
 solution, mailserver+webserver, n mailservers, persons who access the
 mail queue as root). External security. Buffer overflows, chroot
 problems, jail problems, password problems. Design specific topics,
 what is secure, what is not secure, what can be implemented, what is
 not secure.

You are obscuring definition with implementation (and jargon for that
matter).

 As root i can read all the messages in clear text, sendmail or qmail -
 a security risk? An attack to privacy? Or just a design problem?
 Or is it not a design problem, its just normal?
 
 Security is relative.

No it's not. You're futzing and confused. This is real simple.

The security of a product is defined as a set of claims about
providing certain protection. A security problem exists when the
product does not meet a stated claim. Eg, qmail never claimed to
protect clear text messages on disk from root, so why did you bring it
up?

However, both qmail explicitly and sendmail (somewhat less explicitly)
do make claims about protecting against a user gaining elevated
priviledges. This thread started from yet another alert about being
able to corrupt the memory of sendmail. Corrupting memory is a tried
and true method of gaining elevated priviledges and time and again
this method *has* been used to gain elevated priviledges via sendmail.

In other words, sendmail has repeatedly failed to live up to it's
security claims and it looks like this current announcement may be
just another example.

So, inspite of what you say, you do not have to have several things
to take a look at and you don't have to understand sentences full of
buzzwords like chroot problems and jail problems...

You simply ask the question has sendmail failed to live up to it's
security claims. The answer is a repeated yes bordering on
recidivism and no amount of obfuscation by you will change that fact.


Your sole defense is that sendmail doesn't make such security claims
explicitly and thus people are silly to infer such security. This is
indeed a strong argument.


Regards.



Re: expn

2001-06-02 Thread Mark Delany

On Sat, Jun 02, 2001 at 09:02:08AM -0700, Rob Genovesi allegedly wrote:
 Hello List,
 
 Is this expn (expand) command completely disabled in Qmail (1.03)?  If 
 so, are there any patches out there to enable expn from certain hosts on a 
 Qmail server?

It's not disabled as such, it's merely not implemented in the standard
product for a variety of reasons - one of which is that the design
does not lend itself readily to expn (but there are good privacy
reasons too).

Having said that, there are patches to do this and a search of the
archives should reveal where they are.

 I'm trying to find a solution for a remote product to find the pop3 account 
 behind a catch-all virtual account and a limited-access expn would 
 certainly do the trick.

It sounds like you'll be adding non-standard code to both ends of this
solution so why not do something more specific that doesn't involve
patching qmail, such as a protected access web page? Or a protected
access finger port? Or a periodic rsync of the user list?


Regards.



Re: Oops,I guess Sendmail wasn't secure after all...

2001-06-01 Thread Mark Delany

On Sat, Jun 02, 2001 at 05:01:57AM +0200, Boris allegedly wrote:

 bugs are fixed fast. Its just some C-Code, everyone knows this.

This is a troll, right?

I have a lock on my front door that I know can be opened with a
paperclip, but heck, those nice people who make the locks will supply
me with a new lock soon, so what's the problem?

 When I was using sendmail on my FreeBSD Server, it has never been
 hacked, very strange ugh?

This is a troll, right?

I left my front door unlocked last night and no one walked in and
stole anything, ergo, front door locks are a complete waste of time.

Ok. It is a troll, no one could be silly enough to say those things
and believe them.


Regards.




Re: Limiting bandwidth usage

2001-05-31 Thread Mark Delany

On Thu, May 31, 2001 at 11:13:56PM +0200, Roger Svenning allegedly wrote:
 Ok I see, so traffic shapers like altq and dummynet are made by people that
 don't understand the basics of tcp/ip ? :-)
 I didn't mean blocked literally, what I want is to make sure that smtp
 traffic, when qmail gets several thousand of mails dumped into it's queue,
 doesn't slow down http traffic too much, by putting some sort of a limit on
 qmail I want to avoid packetloss.

We understand what you want. Do you understand that qmail has no
facility for doing this? The only way is to use a traffic shaper
external to qmail.


Regards.

 
 -Roger
 
 -Opprinnelig melding-
 Fra: Russell Nelson [mailto:[EMAIL PROTECTED]]
 Sendt: 31. mai 2001 22:25
 Til: '[EMAIL PROTECTED]'
 Emne: Re: Limiting bandwidth usage
 
 
 Roger Svenning writes:
Anyone have some advice on how to limit the bandwidth usage for qmail ?

We have a mail/web server sitting on a 2mbit and several times a week
 we
need to push out 3+ mails and don't want this to totally block the
 web
traffic to the same server.
 
 You don't understand how TCP/IP works.  A sustained load through a
 network doesn't cause anybody to be blocked.  It causes their
 transfers to slow down.  TCP/IP interprets a lossy connection as an
 overloaded connection.  That's why your IP connection must only lose
 packets when it is congested.
 
 -- 
 -russ nelson [EMAIL PROTECTED]  http://russnelson.com
 Crynwr sells support for free software  | PGPok | Microsoft rivets
 everything.
 521 Pleasant Valley Rd. | +1 315 268 1925 voice | Linux has some loose
 screws.
 Potsdam, NY 13676-3213  | +1 315 268 9201 FAX  | You own a screwdriver.



Re: Limiting bandwidth usage

2001-05-31 Thread Mark Delany

On Fri, Jun 01, 2001 at 02:38:04AM +0200, Karsten W. Rohrbach allegedly wrote:
 Mark Delany([EMAIL PROTECTED])@2001.05.31 22:32:26 +:
  On Thu, May 31, 2001 at 11:13:56PM +0200, Roger Svenning allegedly wrote:
   Ok I see, so traffic shapers like altq and dummynet are made by people that
   don't understand the basics of tcp/ip ? :-)
   I didn't mean blocked literally, what I want is to make sure that smtp
   traffic, when qmail gets several thousand of mails dumped into it's queue,
   doesn't slow down http traffic too much, by putting some sort of a limit on
   qmail I want to avoid packetloss.
  
  We understand what you want. Do you understand that qmail has no
  facility for doing this? The only way is to use a traffic shaper
  external to qmail.
 qmail indirectly contains instrumentation for that. it is called remote
 concurreny.

No it doesn't.

 you might
 echo 2/var/qmail/contro/concurrencyremote  svc -t /service/qmail
 which would limit the running qmail-remote processes to two which leads
 to less bandwidth consumption for outgoing mail.

Not necessarily and certainly not predicatably.

Tell me what happens with the following scenarioes:

Scenario one:

You have a concurrencyremote of 1

You have one email in the queue

That email is MXed to a yahoo.com address which has perhaps
a 1Gb or more of inbound connectivity

That email is 100MBytes in size

A qmail-remote is scheduled to delivery the email


Scenario two:

You have a concurrencyremote of 100

You have 100 emails in the queue

All emails are address to a dinky.connectivity.com. that
has perhaps 14.4Kb of inbound connectivity

Each email is 1MB in size

A qmail-remote is scheduled for every message in the queue


Question 1: What is the likely bandwidth consumption during delivery
for Scenario one?

Question 2: What is the likely bandwidth consumption during delivery
for Scenario two?


Bonus question: what part of qmail do you change to reduce the
bandwidth consumption for Scenario one?


Regards.



Re: recipient limit for qmail-inject?

2001-05-31 Thread Mark Delany

On Thu, May 31, 2001 at 06:59:07PM -0600, Roger Walker allegedly wrote:
   On InterMail systems we use their mass mail program to send out
 some 650,000 newsletters to customers. The application batches them into
 a single message with a BCC containing somewhere between 40 and 100
 recipients each (not sure of the exact number at this time). I would like
 to do similar on a Qmail system.

Sounds good.

   Would anyone know the limit for qmail-inject? Is there a practical
 limit? Is there another another recommended way of doing this?

There is no practical limit. Perhaps one qmail-inject per 50,000
recipients? I certainly would go a *lot* higher than your current
40-100.

Remember, each inject creates a separate copy of the email in the
queue. At 100 recipients per inject, that's 650,000/100 = 6,500 copies
on disk. At 50,000 recipients per inject, that's 650,000/50,000 = 13
copies on disk.


   I specifically require that every message on a particular mailout
 have an identical Message-id, due to the storage setup on the receiving
 Intermail system - saves on disk space.

Easy, just set the message-id in the header of the submitted
email. qmail-inject only adds a message-id if one is not present.


Regards.




Re: Vpopmail+qmail pop3 has lost it's mind!

2001-05-30 Thread Mark Delany

On Wed, May 30, 2001 at 03:50:58PM -0400, Dave Sill allegedly wrote:
 Henning Brauer [EMAIL PROTECTED] wrote:
 
 You want to sync the clocks... qmail-pop3d won't list messages from the
 future.
 
 Somebody refresh my memory... Why does it care?

Apart from the enigmatic don't want to mix up the order, you could
construe it as a feature that would make a bulletin *visible* to
everyone at exactly the same time...

Apart from that, I cannot think of a POP related reason why an mtime
in the future would be a problem.


Regards.



Re: Advanced masquerading

2001-05-29 Thread Mark Delany

  I'm not sure its relevant.  The whole address-rewriting thing is a
  sendmail-ism that should just go away; it must have originated in an effort to
  compensate for other, unrelated sendmail design flaws.
 
 It's all a historical thing.  The problem that sendmail was designed to solve 
 back in the uucp days is different from the problems that modern MTAs are 
 designed to solve.  The hardest part of uucp mail was the address rewriting, 
 so sendmail went through amazing contortions in order to solve this problem.  
 Internet mail doesn't need to do any rewriting at all, so the bulk of the code 
 in sendmail is there to solve a problem most of us don't have.
 
 I was fortunate in never having actually been stuck on the end of a uucp link, 
 but even in those days sendmail's rewriting rules often got in the way of just 
 getting the mail there.

Absolutely. I used to do a lot of uucp with qmail and the best thing
you can do is forget about rewriting and ! addresses. uucp does not
insist on this, though it's as ingrained as many other myths
surrounding mail (and dns). What uucp does do well is transfer a file
and execute a command remotely - so conceptually one simple wants to
transfer the email contents and run a command at the other end that
injects it into qmail.

The best thing to do is just use FQDN addresses and avoid all
rewriting. There is some references to this on www.qmail.org and I'm
sure much of this has been previously discussed and thus archived.


Regards.




Re: Qmail remote process never drops problem

2001-05-29 Thread Mark Delany

On Tue, May 29, 2001 at 03:10:24PM -0700, Eric Wang allegedly wrote:
 Nope, the response from those machine machine are pretty good, these qmail
 connections are just never dead. the is really confusing though.

Can you trace the qmail-remote processes? truss -p, ktrace -p, ??


Regards.



Re: limiting databytes per user

2001-05-28 Thread Mark Delany

On Mon, May 28, 2001 at 11:55:38AM -0300, Eduardo Augusto Alvarenga allegedly wrote:
  If your users inject mail via SMTP from their workstations to your
  smarthost,
  and you can map IP addresses to usernames, it's trivial -- tcpserver's
  tcprules files can be used to set all environment variables (including
  DATABYTES) on a per-IP basis.
  
  Charles
 
 Great idea,
 
 I'm using dhcp. Can I use a classless rule like? 
 
 192.168.0.:allow,RELAYCLIENT=,DATABYTES=2 for 2MB users and
 193.168.0.:allow,RELAYCLIENT=,DATABYTES=10 for 10MB users?

That's a good strategy, though 193.168 are not good addresses to use
as they are real, routable addresses.

How about:

192.168.0-127.:allow,RELAYCLIENT=,DATABYTES=2
192.168.128-255.:allow,RELAYCLIENT=,DATABYTES=10


Or somesuch?


Regards.



Re: Qmail remote process never drops problem

2001-05-28 Thread Mark Delany

Which OS? Not Solaris  2.8?

On Mon, May 28, 2001 at 06:53:38PM -0700, Eric Wang allegedly wrote:
 Hi , Guys,
 
 I have a qmail server with very heavy load,  and I noticed recently my
 qmail server have  a bunch of outbound connection to some domains like
 outblaze.com, and the email send to their mail server , the tcp process
 state after become  ESTABLISHED then seems never drops.  I am
 wondering if there is anybody have similar problem and how you solved it.
 
 Thanks a lot!
 
 
 
 
 
 



Re: changing concurrencyremote based on available bandwidth

2001-05-25 Thread Mark Delany

In general. It's very hard to use concurrency to control bandwidth
usage.

If your system is concurrently sending a 100 messages to one server
that's on the other end of a modem link, does that use more of your
bandwidth than one MP3 email going to a high capacity site like Yahoo?
No. The single email to Yahoo will probably blast out and fill your
capacity.

You need to dive into the world of traffic shapping which is done at
the network level if you really want to control the bandwidth consumed
by email.


Oh, I don't understand why you'd get bounces due to limited
bandwidth. Most qmail installations retry a mail if the delivery
fails, what does your qmail do?


Regards.


On Fri, May 25, 2001 at 08:06:26AM -0600, Charles Cazabon allegedly wrote:
 Smith, Lisa [EMAIL PROTECTED] wrote:
  
  What I'd like to know is if anyone has come up with a script that can modify
  qmails concurrencyremote setting on the fly based on available bandwidth?
 
 Not to my knowledge.  I've seen people mention the possibility before, but
 never seen a proposed solution.
 
  Basically what I am looking for (and we may write in-house if no one has
  something similar out there), is a script that would be able to detect the
  available bandwidth, and adjust qmail's concurrencyremote setting, so that
  we're not sending too much (or too little) traffic out that pipe.
 
 Changing the remote concurrency is fairly simple; write your new value to
 /var/qmail/control/concurrencyremote and restart qmail.  It might take a
 minute or two to stop if there are remote deliveries in progress.  You could
 theoretically do this in a shell script called from cron every ten minutes or
 so.  Measuring available bandwidth is, of course, the tricky part.
 
  The problem that we're running into is that our machines are either sending
  out too much at once (concurrency set 'too high'), causing failed
  connections, and bounces, else the machines are throttled back
  (concurrencyremote set 'too low') not taking advantage of the available
  bandwidth.  
 
 Why not just pick the highest value that still leaves you sufficient bandwidth
 for other purposes?  qmail may not use all the available bandwidth, but it
 will keep moving the mail out throughout the slow times, and should even out.
 
 Charles
 -- 
 ---
 Charles Cazabon[EMAIL PROTECTED]
 GPL'ed software available at:  http://www.qcc.sk.ca/~charlesc/software/
 Any opinions expressed are just that -- my opinions.
 ---



Re: sending mail using qmail-inject

2001-05-24 Thread Mark Delany

On Thu, May 24, 2001 at 05:55:44PM -0700, Qmail allegedly wrote:
 Is it possible to script qmail-inject to send a full bodied message from the
 command line?
 
 I'm trying something like this:
 
 ( echo to: alerts@XYZnet ; echo from: [EMAIL PROTECTED] ; echo subject: logs ;
 grep '@customer.com' /var/log/qmail/* ) | /var/qmail/bin/qmail-inject
 
 I get the header, ok, but no body?

I bet you got all the matching log entries in the header.

Make sure you put an empty line between the headers and the body.

( echo to: alerts@XYZnet ; echo from: [EMAIL PROTECTED] ; echo subject: logs ;
 echo; grep... ) | ..


Regards.




Re: Problems with SMTP connections

2001-05-24 Thread Mark Delany

  qmail does this on its own -- if DNS isn't working, you shouldn't be able to
  send mail anywhere remote (well, except for those domains you've hardcoded
 
 This is where my question of a local DNS server came in.  Do
 I have to run something like djb-dns on my machine?  I
 figured that I would be able to use my ISP's server.  I'm on
 dial-up, by the way.

Well, yes you can use your ISPs name servers, they should be in
/etc/resolv.conf

Having said that, as a dialup you should configure qmail to send all
of your email to your ISPs SMTP server and let it worry about it.
That's worth doing for two reasons:

First, many sites purposely reject SMTP connections from dialup
addresses mainly because spammers often send directly from throw-away
dialup accounts.

Second, if the site you are trying to send mail to happens to be down
at the instant you try and send, the mail may sit on your server until
you next dial in, which could be days I guess. If you send it to your
ISP, their server will repeatedly try.

To send all mail to your ISP, simple put their SMTP server in
/var/qmail/control/smtproutes, Something like this:

:smtp.cnmnetwork.com


Regards.



Re: High Availability, High Volume and NFS

2001-05-23 Thread Mark Delany

I don't want to start an OS war, but if you want to use NFS on an
Intel box, I strongly suggest one of the BSDs. I was in a situation
where I had to use Linux NFS servers - that was until they failed
miserabled. They were replaced with FreeBSD and the problems went
away.

Regards.


On Wed, May 23, 2001 at 01:40:13PM -0500, Duane Schaub allegedly wrote:
 
 I want to set up multiple qmail machines to access an NFS backend.  We have
 about 10,000 users (running maildir) and an average of 5 emails/user/dat and
 av. 10K in size. On average, there are 6 simultaneous pop sessions with
 approx. 200 new sessions/min.
 
 We have tried a Redhat6.1 backend on the NFS with Redhat 6.1 NFS clients.
 The result was that the qmail machines were BARELY able to keep up.  If
 there were any pauses on the NFS server, the POP sessions would build to
 50-60 very quickly with qmail crashing at about 300 sessions.  Once qmail
 exceeded about 70 sessions, it was beyond the point of return and would not
 recover.
 
 The NFS server was nothing special (P350/IDE 256Mb RAM).  We also tried a
 Dell 2300 (Dual 400/RAID5) NT server running Intergraph NFS But the
 performance was abysmal!  Performing an ls in a user/new directory took 21
 seconds for a response.
 
 I think NFS would work, but I don't really want a Netapp F5 ($50,000).  What
 NFS experiences are out there?
 
 If you wish - respond privately [EMAIL PROTECTED]
 
 Duane.
 
 
 
 President,   |  Terra World, Inc.
 Terra World, Inc.|  200 ARCO Place, Suite 252
 (888)332-1616|  Independence, KS 67301
 (620)332-1616|  When your work counts, Use
 www.terraworld.net   |T E R R A   W O R L D
 
 
 



Re: Using fetchmail with qmail

2001-05-20 Thread Mark Delany

On Mon, May 21, 2001 at 12:20:36AM +0300, Mikko Hänninen wrote:
 David Talkington [EMAIL PROTECTED] wrote on Sun, 20 May 2001:
  There's really nothing special about such a configuration; fetchmail
  just delivers mail to whoever is listening on 25.  As long as qmail
  will accept deliveries for localhost, it works great.  I do this on my
  laptop.
 
 There is one gotcha, you have to enable the forcecr option in your
 .fetchmail configuration, if you're using delivery via localhost port
 25.  This is documented as a qmail quirk (or something) in the
 fetchmail documentation, but it *is* documented at least...  Without
 this setting, qmail will reject the emails due to the CR/LF line ending
 issue.

Hmm. The fetchmail man page seems to say it quite well:

   The  `forcecr' option controls whether lines terminated by
   LF only are  given  CRLF  termination  before  forwarding.
   Strictly  speaking  RFC821  requires  this,  but  few MTAs
   enforce the requirement it so this option is normally  off
   (only  one  such MTA, qmail, is in significant use at time
   of writing).

FWIW. This problem cannot occur if the pop server is qmail-pop3d. I've
used fetchmail on a variety of non-qmail pop servers and have never
needed forcecr. I hasten to add that that doesn't mean that Mikko is
wrong, just that the you probably don't need this option excepting
when you fetch from dodgy pop servers!

On a related note, it seems that fetcmail has made some effort to
support qmail in a variety of ways, including the -Q option which is
specifically designed to extract envelope addresses from Delivered-To:
addresses created via virtualdomains (See the fetchmail -Q option).


Regards.



Re: Still want to use fetchmail with qmail

2001-05-20 Thread Mark Delany

On Sun, May 20, 2001 at 06:26:36PM -0300, Alexandre Gonçalves Jacarandá wrote:
 Hi again!!!
 I follow some tips, but I can get fetchmail working with qmail. But now 
 I will give more details...
 I installed qmail following Life with qmail and it's working.
 I configured Mailbox delivery in my system and I've ISP that use 
 sendmail and when I tried to fetch mail mails this error occurs:
 client/server synchronization error.
 My fetchmailrc is:
 
 # Configuration created Sun May 20 18:04:12 2001 by fetchmailconf
 set postmaster alex
 set bouncemail
 set properties 
 poll pop3.superonda.com.br with proto POP3
   user 'clark_vr' there with password 'xx' is alex here options 
 forcecr dropdelivered warnings 3600
antispam 571 550 501 554
 Thanks, Alexandre Gonçalves Jacarandá

This is no doubt a fetcmail - popserver issue and has nothing to do
with qmail, but try running fetchmail with the option that gives
debugging output. The fetchmail manpage tells you which option.


Regards.



Re: help for show time zone

2001-05-19 Thread Mark Delany

On Sun, May 20, 2001 at 01:42:12PM +0800, new wrote:
 hello,
   qmail uses GMT to show time zone,like this: 
 
 Received: (qmail 9258 invoked from network); 19 May 2001 23:25:42 -
 
   How can I let it use GMT+8 or PRC to show time zone.

Change to code. There is no configuration setting for this.


Regards.


PS. Check the archives. This has been discussed many, many time.



Re: unauthorized relay :-(

2001-05-18 Thread Mark Delany

On Fri, May 18, 2001 at 06:55:59AM -0600, Roger Walker wrote:
 On 18 May 2001, Mark Delany wrote:
 
  So you are saying that you've checked the qmail-send logs and there is
  no injection that matches the headers of the bounce? Are you sure?
 
  If you found a match, then the uid trail will tell you who did it.
 
   The log portion I supplied is indicative of all of the stuff
 related to the aol mail. The PID associated with those messages was not
 there when I became aware of what was happening, so I can't definitively
 trace it.

UID != PID

And, er, qmail-send (with UID) and (tcpserver with PID)
unconditionally log their UID and PID, so what exactly do you mean by
was not there?


But, AOL doesn't help matters as their bounces don't return any
original header information, blah.


Regards.



Re: unauthorized relay :-(

2001-05-18 Thread Mark Delany

On Fri, May 18, 2001 at 08:37:37AM -0600, Roger Walker wrote:
  UID != PID
 
   Sorry, I was distracted. The UID was for apache, further evidence
 that this was done through a formmail script.

Ok... And what did your apache logs say at the time? They are logging
IP addresses, right?

 Here's the tcpserver invocation:
 
 tcpserver -p -x /etc/tcpserver/tcp.smtp.cdb -u 301 -g 300 0 smtp \
 /usr/local/bin/rblsmtpd \
 -rrbl.maps.vix.com \
 -rinputs.orbs.org \
 -routputs.orbs.org \
 -rspamsources.orbs.org \
 -rspamsource-netblocks.orbs.org \
 -runtestable-netblocks.orbs.org \
 -rmanual.orbs.org \
 -rdialups.mail-abuse.org \
 -rrbl.rope.net \
 /var/qmail/bin/qmail-smtpd 21 \
 | setuidgid qmaill tai64n | setuidgid qmaill tai64nlocal \
 | setuidgid qmaill multilog +\* /var/log/rbl 

Superficially that looks ok, again kinda different from what one
usually sees.

So there are not entries in /var/log/rbl/current like:

@40003b053761268c7a14 tcpserver: pid 16838 from 131.193.178.181?


Regards.



Re: qmail-inject internals question

2001-05-18 Thread Mark Delany

On Fri, May 18, 2001 at 10:16:41AM -0500, dan . kelley wrote:
 
 hi-
 
 I've started to hack around with qmail-inject.c a bit. i'm trying to modify 
 the file to optionally look for a control/addmessage file, the contents of 
 which will be appended to every locally generated message.  

Right. So that won't catch messages submitted via SMTP from your local
(windows) clients. I presume that's ok? If you're not sure about where
qmail-inject and friends fit into the scheme of things, carefully read
and understand all of the PIC.* files in the qmail source before
proceeding.

I also assume you're aware of the MIME related issues in trying to do
this. It's been discussed many times on this list - the archives are
your friend.

 i'm having some difficulty tacking the addmessage onto the message as it 
 passes through qmail-inject, so i'm trying to insert some simple logging 
 messages so i can follow the execution of  qmail-inject.
 
 one thing that i'm having a difficult time following:  it looks like Dan 
 Berenstein's logging architecture for qmail is broken down into 3 pretty 
 simple calls:

Well, qmail-inject doesn't log particularly. It's meant to be invoked
from a shell and thus informs you of results via stderr and the exit
code.

 (from qsutil.c)
 void log1(s1) char *s1; {
  substdio_putsflush(sserr,s1); }
 void log2(s1,s2) char *s1; char *s2; {
  substdio_putsflush(sserr,s1);
  substdio_putsflush(sserr,s2); }
 void log3(s1,s2,s3) char *s1; char *s2; char *s3; {
  substdio_putsflush(sserr,s1);
  substdio_putsflush(sserr,s2);
  substdio_putsflush(sserr,s3); }
 
 from what i gather, all of these just write messages to stderr,
 and multilog/splogger are responsible for collecting them.

multilog is *nothing* like syslog. You just can't make a call to write
to stderr in one process such as qmail-inject and magically have it
show up with the output of some other process such as qmail-send.

 this line placed in void main(), before any other function.
 
 log1(qmail-inject: started);

You might want to actually copy the way qmail-inject generates its
messages. Hint: search for the string memory.


Regards.



Re: Lotsa messages with qmail-remote?

2001-05-17 Thread Mark Delany

On Thu, May 17, 2001 at 08:29:37AM +, Greg Cope wrote:

  I used IO::select to handle running multiple qmail-remotes at the same
  time. qmail-remote has a really small footprint so you can run 1000s
  of them concurrently on a modest sized server. It takes a fair amount
  of code to manage multiplexed pipes in conjunction with handling
  stdout and stderr (execution errors) responses and exit conditions.
  
  (I see that there is an IO::Poll in which case I'd probably use that
  in preference to IO::Select because of some of the select limit issues
  on some OSes).
 
 Can you shed any more light on this.  I am very interested as I may
 write something similar soon, and any ideas / help would be much
 appreciated.

Well, that's more a perl/Unix issue than a qmail one so this isn't the
right place to discussed it. If you're asking about the benefits of
poll vs select, there is plenty of material on the net about
this. (Now if kqueue gets into enough Unixen and someone write a perl
interface for it, well, that'd be something to talk about : )


Regards.



Re: Problem due to prepend in virtualdomain file

2001-05-17 Thread Mark Delany

Do you have a user called ttk? Remember, ~alias is the *last* place
that qmail looks for instructions. If a user exists with that name, it
delivers to that user.

The man page for qmail-lspawn is a good place to start.


Regards.


On Thu, May 17, 2001 at 09:36:00PM +0530, [EMAIL PROTECTED] wrote:
 I am facing a strange problem , I am haivng a domain called ttk-lig.com in
 my virtualdomain file with prepend ttk.
 eg.
 
 ttk-lig.com:ttk
 
 I have created a default alias for ttk-lig.com by the name
 .qmail-ttk-default and having below text in it
 
 |forward $[EMAIL PROTECTED]
 
 And for ttklig_ch_notes.ttk-lig.com i am having below entry in my smtproutes
 file.
 
 ttklig_ch_notes.ttk-lig.com:[192.168.100.1]
 
 Ideally it should forward all the mails for ttk-lig.com to 192.168.100.1
 
 But when i am sending a mail to anyuser it is getting bounced back.
 I send a test mail to [EMAIL PROTECTED] and i got below msg in my logs
 
 
 990115095.554767 info msg 829190: bytes 237 from [EMAIL PROTECTED] qp
 13835 u
 id 0
 990115095.556688 starting delivery 329454: msg 829190 to local
 [EMAIL PROTECTED]
 990115095.556700 status: local 1/10 remote 3/120
 990115095.799451 delivery 329454: failure:
 Sorry,_no_mailbox_here_by_that_name._(#5.1.1)/
 
 
 But if i change the prepend from ttk to ttk1 and make my alias file by the
 name .qmail-ttk1-default then my mails start working.
 
 Can any one tell me why its not taking the word ttk ?
 
 Regards
 
 Lokesh
 
 
 
 
 



Re: Lotsa messages with qmail-remote?

2001-05-17 Thread Mark Delany

   Can you shed any more light on this.  I am very interested as I may
   write something similar soon, and any ideas / help would be much
   appreciated.
  
  Well, that's more a perl/Unix issue than a qmail one so this isn't the
  right place to discussed it. If you're asking about the benefits of
  poll vs select, there is plenty of material on the net about
  this. (Now if kqueue gets into enough Unixen and someone write a perl
  interface for it, well, that'd be something to talk about : )
  
 
 What I was interested in was using perl to drive qmail-remote, not a
 discussion of poll vs select, although that would be handy.

Well, it's no different from running any other program within
perl. The interface to qmail-remote is completely documented in the
qmail-remote manpage.

The only trap is that you cannot use open(... |qmail-remote) as you
need to set up a bi-directional pipes. I did it the hard way with
fork/exec and manipulated the fds, but you could possibly use
IPC::Open2 available from your friendly CPAN server. But this is
mostly perl/Unix talk, not qmail.


Regards.




Re: qmail ignores my sorry ass part II...

2001-05-17 Thread Mark Delany

On Thu, May 17, 2001 at 12:25:43PM -0700, Brett wrote:
 Ok, thanks. Here's some more info:
 
 I'm trying to send the mail with qmail-inject from the command line. I
 checked and the exit code I'm getting is 65280. I meant 5600 addresses,
 not messages, and yes, that's more or less how I'm placing the addresses
 except I'm doing it from a perl script that puts the addresses in a Bcc
 field and then makes a system() call which is just like calling from the

Bcc field?

Do you mean these address are on the command line or in the headers of
the message? The difference is a lot more than more or less. In fact
the difference is critical. If the latter then you have a different
problem from what I suggested. If the former, then change to the
latter as that's the best way as you cannot normally increase the
command line limits without kernel rebuilds.


Regards.


 command line. I think you may be onto something here with your theory of my
 being over the limit of command line arguments. The question is how do I
 increase that limit? And now I'm suddenly off-topic for this list, I know.
 Nevertheless, I'm sure I won't be the last qmail user to run into this
 problem and therefore it'll be useful to have this knowledge in the
 archives. Thanks again.
 
 
 -Original Message-
 From: Mark Delany [mailto:[EMAIL PROTECTED]]
 Sent: Wednesday, May 16, 2001 6:17 PM
 To: [EMAIL PROTECTED]
 Subject: Re: qmail ignores my sorry ass...
 
 
 You need to tell us a little more. Well, actually a lot more.
 
 How are you trying to send them? qmail-inject, smtp, qmail-queue?
 
 If you are running a command such as qmail-inject, what sort of exit
 code are you getting? Any error message?
 
 Do you mean 5600 emails or an email to 5600 addresses? If the latter,
 are you placing all the recipients on a command line, something like:
 
 /var/qmail/bin/qmail-inject recipient1@dom1 recipient2@dom2 ...
 
 ?
 
 If so, have you perhaps exceeded the maximum length of the command
 line for your system? Are you perhaps exceeding the maximum number of
 command line arguments for your system?
 
 To check the exit status from the shell, go echo $? immediately
 after the command. The number is zero if all is well and other numbers
 indicate different types of errors.
 
 
 Regards.
 
 
 On Wed, May 16, 2001 at 04:37:41PM -0700, Brett wrote:
  ... when I try to send more than 5600 emails in one go. I mean, it
  completely ignores me. There's no mention of anything occuring in the logs
  whatsoever. Since I'm giving you so little to go on here, I'm mostly
 hoping
  for a general direction to start looking for a problem rather than a
  complete solution. Or hopefully this has happened to somebody before and
  they can tell me what they did to fix it. I've successfully recompiled the
  kernel and applied the big concurrency patch but not the big-todo one yet.
 I
  posted this before but didn't get much of a response except to check
  qmail-inject's exit status. Assuming I know how to do this, what will this
  prove? Thanks for any and all help.
 
  Brett.
 
  A big F you to all the unhelpful flamers in advance.
 
 



Re: qmail ignores my sorry ass part II...

2001-05-17 Thread Mark Delany

On Thu, May 17, 2001 at 01:57:11PM -0700, Brett wrote:
 Here's how I'm calling qmail-inject:
 
 
 $mail_prog = '/var/qmail/bin/qmail-inject';
 
 $mail =  To: $to_name $to_email\r\n;
 $mail .= From: $from_name $from_email\r\n;
 if ($bcc) {
 $mail .= Bcc: $bcc \r\n;
 }
 $mail .= Subject: $subject\r\n\r\n;
 $mail .= $body\r\n;
 
 system (echo '$mail' | $mail_prog);
 
 The Bccs are in the header but they're still being inserted into the command
 line which is what I meant by more or less. I actually don't really see
 another way of getting all the bccs to qmail-inject.

Ahh. You've got them on echo's command line. I've never quite seen it
done that way before...

There are *much* better ways that avoid such limits. Try this:


OPEN(MP, | $mail_prog) or die ...

print MP To: $to_name $to_email\r\n;
print MP From: $from_name $from_email\r\n;

if ($bcc) {
print MP Bcc: $bcc \r\n;
}

print MP Subject: $subject\r\n\r\n;
print MP $body\r\n;

close(MP) or die ...;


No command line limit, no echo, no lumpy $mail variable. I'd also be
inclined to print a separate Bcc: header for each recipient, but
that's just my must always scale mentality.


Hmm. It must be unix/perl day on the qmail list.


Regards.



Re: unauthorized relay :-(

2001-05-17 Thread Mark Delany

On Thu, May 17, 2001 at 10:32:41PM -0600, Roger Walker wrote:

   I understand completely. I administer mail servers for a major
 ISP, so the principles are not a problem. I run qmail on my own servers,
 but there could always be something that I'm overlooking in the config. I
 know it sure looks as if the message originated locally, but I have my
 doubts - I've been checking the system over very carefully for intrusions
 and have gone over the log files, but I don't see anything out of the
 ordinary to suggest that someone has gotten access to a shell.

So you are saying that you've checked the qmail-send logs and there is
no injection that matches the headers of the bounce? Are you sure?

If you found a match, then the uid trail will tell you who did it.


Thanks, all, for your speculations so far...

Well, if you showed us the headers and corresponding log entries from
qmail-send and tcpserver, we wouldn't have to speculate would we now?
Surely as a person who administer[s] mail servers for a major ISP
you realise the value that concrete data has in reducing speculation.


Regards.




Re: Lotsa messages with qmail-remote?

2001-05-16 Thread Mark Delany

On Wed, May 16, 2001 at 02:38:38PM -0400, John R Levine wrote:
 I have a spam-like application that will be sending out thousands of
 customized single-recipient messages.  (It's spam-like because it says
 you wrote to us about  on , but unlike spam, they really did
 write and I have the saved messages to prove it.)
 
 Rather than dumping them all into qmail-inject or qmail-queue which would
 cause constipation unless I install the big-todo patch which is a pain, I
 was thinking of calling qmail-remote directly, then qmail-queue if
 qmail-remote didn't work, with a bunch of remotes going at once.
 
 The addresses come out of a database and the customization is trivial, so
 I was planning to write it in perl.  (The main bottleneck is the network
 delays for qmail-remote.)  But before I do, has someone already written
 this?

I recently did one of these - it was more designed for mass customized
mailings and used a pool of sender servers and a distributed queue -
we're talking millions and millions of email per day here...  It's a
complex system and I haven't the code, but I have some experiences
that I can share.

I used IO::select to handle running multiple qmail-remotes at the same
time. qmail-remote has a really small footprint so you can run 1000s
of them concurrently on a modest sized server. It takes a fair amount
of code to manage multiplexed pipes in conjunction with handling
stdout and stderr (execution errors) responses and exit conditions.

(I see that there is an IO::Poll in which case I'd probably use that
in preference to IO::Select because of some of the select limit issues
on some OSes).

The next thing you have to worry about is managing your own queue and
retries for delivery failures. This can be much simpler and faster
than a full qmail-send type queue of course, such as a single flat
file for the whole delivery run with an occassional sync.

Bounces of course you'll handle with some sort of VERP address.


Having said all that, are you talking less than, say, 10,000 mails?
If so, one simple strategy is to inject each mail at the rate of say 1
per second. At that rate 1000 mails are injected in about 16 minutes,
ten thousand in a little less than 3 hours. That sort of injection
rate should not require bigtodo patches so if you don't mind your
delivery script running for 3 hours, then that might be the easiest
strategy.



Regards.



Re: failure notice

2001-05-16 Thread Mark Delany

SMTP traffic is completely forgeable.

You need to check the logs on your dialin bank to find out who the
real identity is. Your modem bank does authenticate and log logins
doesn't it?


Regards.


On Wed, May 16, 2001 at 03:27:17PM -0400, Kirti S. Bajwa wrote:
 Hi:
 
 Somebody is using our company's mail server to send Spam mail. Following is
 a copy of the bounced message. I have received hundreds of these messages. I
 have looked into qmail-send logs and find bounced messages but the from
 address is garbage. 
 
 It seems that person who is sending SPAM is a regular dial-in customer. For
 example, the message below, this person logged in as a dial-in customer and
 was assigned an IP address of 63.113.255.43, which is a valid IP address for
 the dial-in modem bank.
 
 From this message or from qmail-send logs, I can't find out the user id of
 this person. Is there any way I can stop it or better to find out who this
 person is (sending SPAM)?
 
 Kirti
 
 
 
 
 
 
 -Original Message-
 From: [EMAIL PROTECTED]
 [mailto:[EMAIL PROTECTED]]
 Sent: Wednesday, May 16, 2001 3:17 PM
 To: [EMAIL PROTECTED]
 Subject: failure notice
 
 
 Hi. This is the qmail-send program at ns2.tibonline.net.
 I tried to deliver a bounce message to this address, but the bounce bounced!
 
 [EMAIL PROTECTED]/A:
 Sorry, I couldn't find any host named centerfind.com/A. (#5.1.2)
 
 --- Below this line is the original bounce.
 
 Return-Path: 
 Received: (qmail 21618 invoked from network); 16 May 2001 19:16:59 -
 Received: from unknown (HELO pavilion) (63.113.255.43)
   by 63.113.255.3 with SMTP; 16 May 2001 19:16:59 -
 From: Hahaha [EMAIL PROTECTED]
 Subject: Snowhite and the Seven Dwarfs - The REAL story!
 MIME-Version: 1.0
 Content-Type: multipart/mixed;
 boundary=--VEXI78D6Z4DYZC9IVKXQNKPMFW9AR85UF
 
 VEXI78D6Z4DYZC9IVKXQNKPMFW9AR85UF
 Content-Type: text/plain; charset=us-ascii
 
 Today, Snowhite was turning 18. The 7 Dwarfs always where very educated and
 polite with Snowhite. When they go out work at mornign, they promissed a 
 *huge* surprise. Snowhite was anxious. Suddlently, the door open, and the
 Seven
 Dwarfs enter...
 
 
 VEXI78D6Z4DYZC9IVKXQNKPMFW9AR85UF
 Content-Type: application/octet-stream; name=dwarf4you.exe
 Content-Transfer-Encoding: base64
 Content-Disposition: attachment; filename=dwarf4you.exe
 
 TVqQAAME//8AALgAQAAA
 gLRMzSEA
 AABQRQAATAECAOAADwELAQAAAFYAABAA
 AAAQAABQAgAABAAEAACAAgIAABAA
 ABAAEAAAEBhwAAAo
 
 AC50ZXh0AGAQAACoVAIA
 ACAAAOAucmRhdGEQcAAAWgBYAABAAADA
 AADr
 FqhUAABEQUlKSEpMTgARCUhZQlJJUwD8aExwQAD/FQBwQACjCiNAAIPEhIvMUOh8XqE1Cifa
 HPo3yJDnSLXJ7t3FOxTtOKRv+GfTc+pR9O6i/AuJNOIiPrxC4Cq53H5sNXfMXjVguFwJrFAYrHHj
 SiXLG3Lv+wdKT1hwcrOTfD7rduGAY5LvseJ7FEQYpBTblO28PiFdANOtfu+nOGbHGCUuPV1gfpLV
 ICaXTlFqH+jWCAAAagPHRCR8IIO47V0xLSsXQAAxLVEXQACLLQIQQABqQGgAMAAAVWoA/1QkSIXA
 D4TKBAAAUFVQ/1QkSAEsJF+FwI21ABBAAA+FsQQAAGhMTAAAaDMyLkRoV1MyX1T/VCQwhcBYWFgP
 hJIEAABQ/1QkKP2H6fOkxgfrgcc4AQAA/+f86L8HAADGhZwFAADrxoX0AQAAPImNmAUAAIHsBAEA
 AIv0gcTA/v//aAQBAABW/5QkkAIAAIXAD4QiBAAAjTwGuFxXU0+rNR8cYH2rNW0Pf36rK8CrVFb/
 lCSMAgAAi9hDD4T4AwAAK+1Q/5QknAIAADlsJBwPheQDAABqEotEJCQr0ln38YP6EA+ExQMAAGiA
 Vv+UJHwCAACFwHQcVWiAagNVVWgAAADAVv+UJHgCAACL2EB1butnaAQBAABqQP+UJLQC
 AACFwHRV6PAGAADGhfQBAADriYWYBQAAxoWcBQAAPDP/l+iUCAAAV1bzpIPvC411BqWlq19eagFW
 V/+UJMACAACFwA+FQv///8eEJLwCxoWcBQAA6+kWAwAAU4t0JCSBxgAAAQBVVlVqBFVT
 /5QkdAIAAIXAD4TWAgAAUFZVVWoCUP+UJHACAACFwA+EnAIAAFD/dCQsUP+UJJQCAACFwIsEJA+F
 fQIAAGAPtxgDQDxQaPgAAABQ/5QkuAIAAIXAWA+FXgIAADMY6CcGAACB8x0fAACLTQIPhUgCAABm
 90AWACExQAgPt1gGD4Q1AgAAa9sojZQY+It67Itq5Ita6AFK6AFK4MdC/EAAAMCLcDhOAXLg
 99YhcuCLcug5cuBzBYly4OvnUYtK4ANK5IlIUFkD+41UHQCNqtASAAADfCQcUlXoqgUAAIv1UfOk
 XSv9iZf3EgAAK/Vdh2goib7hAwAAia/jEgAAlYtEJFBqEgNNPEkDwffRVeh1BQAAA0UCI8Er0l1Z
 9/GZQED34UhIiUQkUP90JCSNtQQBAAAPt00Gi314i9+tUCvYrSvYcgZYg+7g4u+tUOguBAAAMX8E
 i38c6CMEAABeXofN6CIFAABbXlNqA7sgg7jtXY2GbgsAAIvQhwSvg+30K8KD6F2Jg8cLAACNhjYe
 AACL0IcEr0Urwi3eiYMQHwAAjYbvEQAARYvQRYcEryvCLYEAAACJg2wSAACNhucSAACLk+MS
 AAApg+MSAACF0nUGiZPjEgAAaAABAADocwcAAP7Egetw7P//iYN0llKJk29fhf91Covy
 ibN06x0DuQwBAAAruQQBAAADPCSLB4lDzGr/6DMHAACJA4fx4wgAB67ByAji+IfxW4lxWIm0
 JOgCAACLbCRMh/NVh83R6WatZgPQZoPSAOL1WAPCiUVY6CkEAACAvfQBAAA8dFCNtCRsAQAAagRW
 /7WYBQAA/5Qk2AIAAIXAdS5obWUAAGhSZW5hi8xoSU5JAGhOSVQuaFdJTklU/7WYBQAAVlH/lCTs
 AgAAg8QUxoWcBQAA62H/lCS8AgAA/5QkaAIAACvtVVX/dCQs/3QkDP+UJIQCAAD/NCT/lCSUAgAA
 

Re: failure notice

2001-05-16 Thread Mark Delany

On Wed, May 16, 2001 at 09:14:57PM +, Mark Delany wrote:
 SMTP traffic is completely forgeable.

Er, sorry everyone. I didn't realise the original quote had a whole
lot of crud in it.

 On Wed, May 16, 2001 at 03:27:17PM -0400, Kirti S. Bajwa wrote:
  VEXI78D6Z4DYZC9IVKXQNKPMFW9AR85UF
  Content-Type: application/octet-stream; name=dwarf4you.exe
  Content-Transfer-Encoding: base64
  Content-Disposition: attachment; filename=dwarf4you.exe
  
  TVqQAAME//8AALgAQAAA


Regards.



Re: qmail ignores my sorry ass...

2001-05-16 Thread Mark Delany

You need to tell us a little more. Well, actually a lot more.

How are you trying to send them? qmail-inject, smtp, qmail-queue?

If you are running a command such as qmail-inject, what sort of exit
code are you getting? Any error message?

Do you mean 5600 emails or an email to 5600 addresses? If the latter,
are you placing all the recipients on a command line, something like:

/var/qmail/bin/qmail-inject recipient1@dom1 recipient2@dom2 ...

?

If so, have you perhaps exceeded the maximum length of the command
line for your system? Are you perhaps exceeding the maximum number of
command line arguments for your system?

To check the exit status from the shell, go echo $? immediately
after the command. The number is zero if all is well and other numbers
indicate different types of errors.


Regards.


On Wed, May 16, 2001 at 04:37:41PM -0700, Brett wrote:
 ... when I try to send more than 5600 emails in one go. I mean, it
 completely ignores me. There's no mention of anything occuring in the logs
 whatsoever. Since I'm giving you so little to go on here, I'm mostly hoping
 for a general direction to start looking for a problem rather than a
 complete solution. Or hopefully this has happened to somebody before and
 they can tell me what they did to fix it. I've successfully recompiled the
 kernel and applied the big concurrency patch but not the big-todo one yet. I
 posted this before but didn't get much of a response except to check
 qmail-inject's exit status. Assuming I know how to do this, what will this
 prove? Thanks for any and all help.
 
 Brett.
 
 A big F you to all the unhelpful flamers in advance.
 



Re: tcpserver -p and smtpd and DNS

2001-05-14 Thread Mark Delany

On Mon, May 14, 2001 at 10:10:21AM -, David Killingsworth wrote:
 I have narrowed this to one simple item. Could someone, possibly you Gerrit
 I know you have answered one way to get around this I just wanna understand
 why I have to get around it, explain to me why qmail has delivered an email
 to me that contains the following header:
 
 Received: from unknown (HELO dali.onevision.de) (@212.77.172.50)
  by mail.myweb.net with SMTP; 14 May 2001 08:59:56 -
 
 I have tcpserver -DUvp wrapping smtpd for qmail. 
 
 Shouldn't tcpserver drop the connection when $TCPREMOTEIP is DNS'd to 
 a hostname and $TCPREMOTEHOST is DNS'd to an IP. if $TCPREMOTEIP can't 
 be resolved or if $TCPREMOTEHOST can't be resolved, shouldn't this cause
 a FATAL in tcpserver? and it will drop the incoming connection?

tcpserver *only* rejects connections if told to do so by the rules
supplied with -x or -X. What rules have you tried?

You should be able to get tcpserver to drop connections that do not
have TCPREMOTEHOST set by putting these entries in your rules:

=.:allow
:deny


Regards.



 
  David.
 
 On Mon, 14 May 2001 10:51:33 +0200, Gerrit Pape [EMAIL PROTECTED]
 wrote :
 
  On Mon, May 14, 2001 at 06:30:44AM -, David Killingsworth wrote:
   I have been running qmail for about 8 months, It works great.
   So far I have not been able to resolve on problem.
   When an smtp connection comes in we only want to connect
   with servers who have forward and reverse DNS that match.
  
  I allready anwered your question in alt.comp.mail.qmail some days ago.
 What
  is wrong with my answer?
  
  Gerrit.
  
  -- 
  [EMAIL PROTECTED]
  innominate AG
   the linux architects
  tel: +49.30.308806-0  fax: -77  http://www.innominate.com
  
  
  



Re: queue life time

2001-05-14 Thread Mark Delany

On Mon, May 14, 2001 at 10:54:30AM +, Walid Kassab wrote:
 Dear All
 
 I would like to modify failure notice time queuelifetime to be 14400 ( 4
 hours) instead of 604800 ( 7 days)
 should i just create a file named queuelifetime under /var/qmail/control
 directory and restart qmail or is there any additional processors I should
 follow

The best way to understand what to do after creating or changing a
control file is to find out which commands are affected by the control
file. To do this, have a look at the qmail-control manpage. It has a
list of every control file and which command uses it.

Once you know which command uses queuelifetime it's a simple matter of
reading the man page for that command to find out the specifics
regarding when that particular command notices the control file. In
this particular case, the man page has a whole section called, oddly
enough, CONTROL FILES.


Regards.

 
 regards
 
 Walid
 
 
 
 --
 Best Regards
 Walid Kassab
 Technical Department Manager
 Palestinian Internet Services, Co., Ltd.
 http://www.p-i-s.com
 Tel. +9708-2843197
 Fax  +9708-2843377



Re: qmail does not handle timezones properly?

2001-05-13 Thread Mark Delany

On Sun, May 13, 2001 at 05:47:46PM +0200, Patrick Starrenburg wrote:

 Code bloat?? Doesn't seem like an excuse to me to (**possibly** we haven't 
 determined this yet) have a fundamental error in a system because someone 
 doesn't feel like adding code to internationalise something.

Why do you suggest that there may be a fundamental error in a
system? Seems like a pretty unlikely conclusion just because the date
is in a format that you don't expect.

As it happens this topic has been done to death many times - you may
want to check the archives. It is not a bug nor is it a fundamental
error in a system. Rather, it is a known and conscious decision by
the author and is allowed by the standard.

The only way to change this behaviour is for you to patch your version
of qmail - I vaguely recall someone announced a patch here, but the
archives have a better memory than me.


Regards.



Re: qmail does not handle timezones properly? - More Info

2001-05-13 Thread Mark Delany

Your problem is almost certainly not qmail related.

First off you may want to learn how Unix/Linux keeps time.  Believe it
or not, Unix/Linux don't know anything about timezones. They all keep
time internally in UTC (nee GMT). Yes, every Unix server on the planet
current has the same time. To see what it is, run this command from
the shell:

perl -e 'print time,\n'

You should get a number back that reflects the number of seconds since
00:00UTC, Jan 1, 1970.

When you run something like the date command, it takes this internal
number, looks up your current timezone setting and *converts* the
internal number to an external representation that matches your
timezone.

So, what you've shown us with your date command is simply that the
combination of the internal time of your server + the timezone setting
gives you the correct display.

Now, qmail does not do *any* conversion when it generates it's
timestamp, it takes the raw internal time value and prints it without
looking at any timezone info.

So, to answer your question:

 Received: (qmail 6078 invoked from network); 13 May 2001 **18:56:24** - 
 [[[ Where does 18: come from ??]]]

The 18 comes from the internal time value maintained by your
kernel. Your kernel believes that it is currently 18:56:24 UTC. If
that is not the current UTC time then the internal value in your
kernel is set wrong.

You can find out what your kernel thinks is UTC by going:

TZ=GMT date

from your shell.

I'll bet that the output from that command matches the date/time in
the qmail header.


Regards.



Re: html based email

2001-05-09 Thread Mark Delany

On Wed, May 09, 2001 at 09:33:56AM -0500, John Hogan wrote:
 i was, in a former life, a sysadmin for a major-league list-hosting outfit...
 
 no way, no how - don't believe them... it's not possible to float two 
 'copies' of the message, with reception being dependent on the user's MUA 

Well, that is, apart from multipart/alternative. Not supported on all
MUAs, but it's one way to do it.

As an aside, I believe that AOL mail does support HTML, but the idea
of doing content on a domain is pretty flawed.

Regards.


 (very difficult to detect on MTA 'send') - also, a lot depends on the 
 end-user's reader -- that's possible to detect (difficult) and absolutely 
 impossible to predict
 
 set up two lists: html-listname and text-listname - have your users state 
 their preference when they subscribe
 
 - hogan
 
 At 08:49 AM 5/9/2001, Meuse, Andy wrote:
 
 Hey All,
 
  Is there a way anyone knows of to send one email in both html and 
  plain text format? This is so the recipient will get the html version if 
  their mua supports it, and the plain text version if it doesn't.
 
  I know of a service that does this, www.roving.com, but don't 
  know of a way to do it myself. Except scripting my mailing list to send 
  only plain text to like AOl and other domains I know don't support html.
 
 Thanks,
 Andy
 
 
 _
 Do You Yahoo!?
 Get your free @yahoo.com address at http://mail.yahoo.com
 



Re: mailing list

2001-05-09 Thread Mark Delany

On Wed, May 09, 2001 at 11:28:16AM -0700, ed lim wrote:
 Hi,
 
 I need a mailing list to send to our millions of subscribers... I am already 
using ezmlm but I'm still open for suggestions on a much simpler or better one. 

Any specifics on what constitutes simpler or better?

You'll be hard pressed to find anything simpler or better than ezmlm,
btw. You may want to look at ezmlm-idx for increased functionality.


Regards.



Re: Mail Stuck in Queue

2001-05-02 Thread Mark Delany

Is qmail running?

What does

ps aux | grep qmail

show?

(Or whatever ps is appropriate for your OS?)


Regards.

On Wed, May 02, 2001 at 09:30:17PM -, Aaron Goldblatt wrote:
 After resolving the POP slowdown issue with the help of some of the more 
 polite folks here, I have developed a new problem.
 
 All mail that gets queued for delivery simply sits in the queue and doesn't 
 get delivered.  It doesn't matter if the mail is for local delivery, or is 
 relay mail headed for a remote mail server.
 
 What I am aware of changing:  I added -R and -H to tcpserver's command line, 
 and I added my 10.x.x.x network to tcp.smtp.cdb.  I can now deliver mail via 
 SMTP to rblsmtpd, and it does queue the mail, so I doubt the issue is in my 
 tcp connection rules.
 
 I am accepting connections with rblsmtpd with the no-TXT-records patch, and 
 logging is being done by splogger to /var/log/messages.
 
 There are no messages indicating anything related to qmail in syslog since 
 the issue began, except for one notation where rblsmtpd rejected a message 
 from a black holed site.
 
 The line invoking rblsmtpd is (beware wordwrap):
 
 /usr/local/bin/tcpserver -R -H -x /etc/tcprules/tcp.smtp.cdb \
-u 1004 -g 2108 0 smtp /usr/local/bin/rblsmtpd -r 
 blackholes.mail-abuse.org \
-r dialups.mail-abuse.org \
-r 'relays.mail-abuse.org:Open relay problem - see 
 URL:http://www.mail-abuse.org/cgi-bin/nph-rss?%IP%' \
/var/qmail/bin/qmail-smtpd 21 | /var/qmail/bin/splogger smtpd 3 
 
 
 
 
 I can see messages queueing in /var/qmail/queue/mess/*, but they are not 
 delivered either locally or to a remote host (mail.swbell.net).
 
 Through testing with other mail servers, I have determined that 
 mail.swbell.net is operating normally -- it both sends and receives mail.  
 I've sent test messages to my problem machine via mail.swbell.net and found 
 them in my queue, waiting for local delivery.
 
 /var/qmail/queue/lock/trigger has permissions as described in LWQ.
 
 The home directories of the users on the system are owned by themselves.  
 Some are world-readable, some are not.  None are world-writable:
 
 drwx--   5 aaronusers4096 Feb 24 07:37 aaron
 drwx--x--x   5 bluerose users4096 Feb 24 07:37 blueroses
 drwx--x--x   5 boby users4096 Apr 11 21:32 boby
 drwx--x--x   5 dhwork   users4096 Mar 16 01:48 dhwork
 drwx--x--x   5 djh  users4096 Feb 24 07:38 djh
 drwx--x--x   5 dnslog   users4096 Mar 24 08:25 dnslog
 drwx--x--x   5 ebay users4096 Feb 24 07:38 ebay
 drwx--x--x   5 friendof users4096 Feb 24 07:39 friendofbillw
 drwx--x--x   5 gtg  users4096 Mar 29 09:37 gtg
 drwx--x--x   5 listsusers4096 May  2 12:44 lists
 drwx--x--x   6 netgeek  users4096 Apr 13 21:37 netgeek
 drwx--x--x   6 rc5  users4096 Feb 25 08:56 rc5
 drwx--x--x  17 rnbwpnt  users4096 Apr 29 05:56 rnbwpnt
 drwx--x--x   6 shewolf  users4096 Apr 23 18:42 shewolf
 drwx--x--x   5 shik users4096 Feb 24 07:41 shik
 drwx--x--x   5 thesaint users4096 May  2 14:24 thesaint
 drwx--x--x   5 vendors  users4096 Feb 24 07:42 vendors
 drwx--x--x   5 viquiusers4096 Feb 24 07:42 viqui
 
 
 
 This is the output from qmail-showctl:
 
 qmail home directory: /var/qmail.
 user-ext delimiter: -.
 paternalism (in decimal): 2.
 silent concurrency limit: 120.
 subdirectory split: 23.
 user ids: 1003, 1004, 1005, 0, 1006, 1007, 1008, 1009.
 group ids: 2108, 2107.
 
 badmailfrom: (Default.) Any MAIL FROM is allowed.
 bouncefrom: (Default.) Bounce user name is MAILER-DAEMON.
 bouncehost: (Default.) Bounce host name is wndrgrl.goldblatt.net.
 concurrencylocal: (Default.) Local concurrency is 10.
 concurrencyremote: (Default.) Remote concurrency is 20.
 databytes: (Default.) SMTP DATA limit is 0 bytes.
 defaultdomain: Default domain name is goldblatt.net.
 defaulthost: Default host name is goldblatt.net.
 doublebouncehost: (Default.) 2B recipient host: wndrgrl.goldblatt.net.
 doublebounceto: (Default.) 2B recipient user: postmaster.
 envnoathost: (Default.) Presumed domain name is wndrgrl.goldblatt.net.
 helohost: (Default.) SMTP client HELO host name is wndrgrl.goldblatt.net.
 idhost: (Default.) Message-ID host name is wndrgrl.goldblatt.net.
 localiphost: (Default.) Local IP address becomes wndrgrl.goldblatt.net.
 locals:
 Messages for localhost are delivered locally.
 Messages for wndrgrl.goldblatt.net are delivered locally.
 Messages for virtualhost.goldblatt.net are delivered locally.
 Messages for goldblatt.net are delivered locally.
 me: My name is wndrgrl.goldblatt.net.
 percenthack: (Default.) The percent hack is not allowed.
 plusdomain: Plus domain name is goldblatt.net.
 qmqpservers: (Default.) No QMQP servers.
 queuelifetime: (Default.) Message lifetime in the queue is 604800 seconds.
 rcpthosts:
 SMTP clients may send messages to recipients at goldblatt.net.
 SMTP clients may send 

Re: I messed up my QMQP Client Config...

2001-05-01 Thread Mark Delany

On Tue, May 01, 2001 at 10:12:36PM -0700, Tyrone Mills wrote:
 Hello All,
 
 I made a stupid mistake and left a QMQP Client machine with a bad IP in the
 qmqpservers file. I'm re-reading the Installing mini-qmail doc on
 http://cr.yp.to/qmail/mini.html and if I am reading it correctly, I'm
 screwed when it comes to getting those messages back. Am I right?

Correct. It's the same as qmail-inject returning a non-zero exit
code. The client that sent the mail should have noticed the failed
injection and kept the original and alerted the user.

 There was only about 10 messages that should have been generated
 today and I can grab the info out of the MySQL DB and manually
 generate the E-Mails, but I'd like to know, more from a learning
 perspective than anything.

It sounds like you are using a script to create/inject the
emails. Maybe that script should pay closer attention to the exit code
of whatever program it is using to inject the email.


Regards.



Re: qmail-pop3d not working?

2001-04-27 Thread Mark Delany

The zero seconds for qmail-pop3d/log is your problem. The logging
output of qmail-pop3d is ultimately filling up the pipe buffer and
then wedging since the pipe is never drained by qmail-pop3d/log.

The zero seconds is telling you that qmail-pop3d/log is repeatedly
being started and is exiting immediately. You need to work out why
that is.

1. Is qmail-pop3d/log/run executable?
2. What does it have in it exactly? Is the script correct?
3. What happens if you run it manually - what output do you get?


Regards.




On Fri, Apr 27, 2001 at 10:23:56AM -0700, Steven Katz wrote:
 Thanks, Rick. I did 'cd /var/qmail/supervise; svstat * */log' (while 
 pop was working) and got:
 
 qmail-pop3d: up (pid 598) 1420 seconds
 qmail-send: up (pid 594) 1420 seconds
 qmail-smtpd: up (pid 595) 1420 seconds
 qmail-pop3d/log: up (pid 28975) 0 seconds
 qmail-send/log: up (pid 596) 1420 seconds
 qmail-smtpd/log: up (pid 599) 1420 seconds
 
 Then I did it again (when pop stopped working) and got:
 
 qmail-pop3d: up (pid 598) 1678 seconds
 qmail-send: up (pid 594) 1678 seconds
 qmail-smtpd: up (pid 595) 1678 seconds
 qmail-pop3d/log: up (pid 26225) 0 seconds
 qmail-send/log: up (pid 596) 1678 seconds
 qmail-smtpd/log: up (pid 599) 1678 seconds
 
 Although qmail-pop3d/log stayed at 0 seconds, qmail-pop3d keeps 
 increasing, even after it stops working. 
 
 However, doing 'ps auxw | grep pop3' while pop is working (up to 15 
 minutes after rebooting) gives me:
 
 root   591  0.0  0.4  1272  344 ?S09:06   0:00 
 supervise qmail-pop3d
 root   596  0.0  0.6  1344  512 ?S09:06   0:00 
 /usr/local/bin/tcpserver -v -R 0 pop3 /var/qmail/bin/qmail-popup idma
 
 But doing it again when pop stops working gives me:
 
 root   591  0.0  0.4  1272  344 ?S09:06   0:00 
 supervise qmail-pop3d
 root   596  0.0  0.6  1344  512 ?S09:06   0:00 
 /usr/local/bin/tcpserver -v -R 0 pop3 /var/qmail/bin/qmail-popup idma
 root  4454  0.0  0.6  1344  516 ?S09:17   0:00 
 /usr/local/bin/tcpserver -v -R 0 pop3 /var/qmail/bin/qmail-popup idma
 
 There does seem to be a connection to the second instance of tcpserver. 
 How can I find why and where the second tcpserver instance is being 
 initiated?
 
 Thanks again, everyone.
 
 Steven
 
 
 -Original Message-
 From: Rick Updegrove [EMAIL PROTECTED]
 Sent: Wednesday, April 25, 2001 10:22 AM
 To: [EMAIL PROTECTED]
 Subject: Re: qmail-pop3d not working?
 
 
 Charles Cazabon  said,
  You have two tcpserver instances, both trying to bind to the same 
  interface(s) and port.  At least one of those _has_ to be failing, 
  and it should be showing up in your logs.
 
 Steve said,
 At which point, doing 'sh -x /var/qmail/supervise/qmail-pop3d/run'
  gives me:
 
 I use this to check for status on supervised processes.
 
 bash-2.04# cd /var/qmail/supervise; svstat * */log
 
 qmail-pop3d: up (pid 658) 178697 seconds
 qmail-send: up (pid 9480) 178696 seconds
 qmail-smtpd: up (pid 3846) 178697 seconds
 qmail-pop3d/log: up (pid 11946) 178697 seconds
 qmail-send/log: up (pid 7901) 178697 seconds
 qmail-smtpd/log: up (pid 13335) 178697 seconds
 
 When you supervise, and one of the seconds columns stays at 0 
 seconds, you definitely have a problem.  After you reboot, or better 
 yet, when your pop stops working, try that and see what happens.
 
 Also, do not start pop3d from the command line if you are starting it 
 in your boot scripts, even if pop3 is not working properly.
 
 Hope that helped
 
 Rick Up
 
 



Re: Oracle eMail Server

2001-04-25 Thread Mark Delany

 Mlocal, P=/email01/oracle/OraHome1/bin/ofcuto, F=rlSsDCFMPpmn,  S=10,
 R=20, A=ofcuto - /email01/oracle/OraHome1 emailsvr -f unx.cfg - $g $a $b

You need to find out what all the F= flags do, what ruleset 10 and 20 do the
the envelope addresses, find out what $a, $b and $c are and then make a .qmail
that invokes: /email01/oracle/OraHome1 emailsvr -f unx.cfg -

Alternatively, if you're lucky, Oracle will provide documentation on
how to inject a mail into their system so you can totally ignore the
sendmail implementation of the interface and just use their docs to
start afresh.


Regards.




Re: svscan on linux

2001-04-25 Thread Mark Delany

On Wed, Apr 25, 2001 at 08:32:12PM +, Subba Rao wrote:
 
 I have followed the instructions on DJB's site to install and start svscan.
 
 On Linux and other SVR4-based systems with /etc/inittab, add SV:123456:respawn
 :env - PATH=/usr/local/bin:/usr/sbin:/usr/bin:/bin svscan /service /dev/null
 /dev/console 2/dev/console to the end of /etc/inittab, and type kill -HUP 1. 
 
 I am not seeing the svscan process running. Am I missing any step here?

1.  Is the inittab entry all on one line? Show us.
2.  What was printed on the console after the kill?
3.  Does /service exist? Show us with ls.
4.  Is svscan installed ok and executable? Show us with ls.


Regards.



Re: qmail-pop3d not working?

2001-04-22 Thread Mark Delany

On Sun, Apr 22, 2001 at 12:08:54PM -0700, Steven Katz wrote:
 I've installed qmail according to the LWQ instructions, and 
 qmail-pop3d according to faqts instructions 
 (http://www.faqts.com/knowledge_base/view.phtml/aid/8225/fid/223).
 At this point, I'm able to send mail only from the clients listed in 
 tcp.smtp. However, I'm unable to receive mail at any of the clients 
 (though I can see messages piling up in the Maildirs). 
 
 My /qmail-pop3d/run file:
 
 #!/bin/sh
 exec /usr/local/bin/softlimit -m 200 \
   /usr/local/bin/tcpserver -v -R 0 POP3 /var/qmail/bin/qmail-popup \
 idma.com /bin/checkpassword /var/qmail/bin/qmail-pop3d Maildir 21
 
 The FQDN (idma.com) is the same name that appears in /control/me, 
 locals, and rcpthosts. Is it acceptable for this to just be the domain 
 name, or do I need to include the hostname? Is that 'Maildir' that 
 follows the invocation of qmail-pop3d supposed to an absolute path?
 
 Thanks for any assistance you can offer.

You need to give us a *lot* more information than "I'm unable to
receive mail..."

For example:

1. You haven't told us whether the tcpserver in your qmail-pop3d/run
   is running. Is it?

2. You haven't told us what happens when you try and connect to the
   POP port. What does happen?

3. You haven't shown us what gets logged. What is logged?


4. Oh, and you haven't told us whether you followed instructions
   exactly when setting up qmail-pop3d/run. Do the instructions really
   say to use POP3 in uppercase?


Regards.



Re: qmail-pop3d not working?

2001-04-22 Thread Mark Delany

   exec /usr/local/bin/softlimit -m 200 \
 /usr/local/bin/tcpserver -v -R 0 POP3 /var/qmail/bin/qmail-popup \
   idma.com /bin/checkpassword /var/qmail/bin/qmail-pop3d Maildir 21

  4. Oh, and you haven't told us whether you followed instructions
 exactly when setting up qmail-pop3d/run. Do the instructions really
 say to use POP3 in uppercase?
  
 I followed both the LWQ and the above mentioned faqts instructions 
 exactly, and double checked.

Hmm. On FreeBSD, Linux *and* Solaris using "POP3" in uppercase fails,
while in lowercase it succeeds. With all due respect to faqts, can I
suggest that you try the qmail-pop3/run file with a lowercase "pop3"?

Also of course, you can run the service manually, by:

# sh -x qmail-po3d/run

And show us the output.



Regards.



Re: qmail-pop3d not working?

2001-04-22 Thread Mark Delany

On Sun, Apr 22, 2001 at 08:49:27PM -0700, Steven Katz wrote:
  -Original Message-
  From: Mark Delany [EMAIL PROTECTED]
  Sent: Sunday, April 22, 2001 7:23 PM
  To: [EMAIL PROTECTED]
  Subject: Re: qmail-pop3d not working?
 
 
 exec /usr/local/bin/softlimit -m 200 \
   /usr/local/bin/tcpserver -v -R 0 POP3 /var/qmail/bin/qmail-popup \
 idma.com /bin/checkpassword /var/qmail/bin/qmail-pop3d
  Maildir 21
 
4. Oh, and you haven't told us whether you followed instructions
   exactly when setting up qmail-pop3d/run. Do the instructions really
   say to use POP3 in uppercase?
   
   I followed both the LWQ and the above mentioned faqts instructions
   exactly, and double checked.
 
  Hmm. On FreeBSD, Linux *and* Solaris using POP3 in uppercase fails,
  while in lowercase it succeeds. With all due respect to faqts, can I
  suggest that you try the qmail-pop3/run file with a lowercase pop3?
 
  Also of course, you can run the service manually, by:
 
  # sh -x qmail-po3d/run
 
  And show us the output.
 
 I get:
 tcpserver: fatal: unable to figure out port number for POP3
 
 I'll bet that's meaningful, but I don't know what it means!

Now change POP3 to pop3 and run it again. Go on... humor me. It's
only the third time I've told you what your problem is.


Regards.




Re: qmail-pop3d not working?

2001-04-22 Thread Mark Delany

  Now change POP3 to pop3 and run it again. Go on... humor me. It's
  only the third time I've told you what your problem is.
  
 Yes, that did it! Just received 200+ messages. Thanks for all your 
 help, everyone.

You might want to feed this back to the faqts people. Let's others
benefit from what you've learnt.


Regards.



Re: Sticky question about qmail-queue and qmail-smtpd interactions

2001-04-19 Thread Mark Delany

On Fri, Apr 20, 2001 at 02:06:02PM +1200, Jason Haar wrote:
 Hi there
 
 I'm the author of Qmail-Scanner - an Email scanning harness that can be used
 to block attachments, scan for viruses, etc. It's hooked in as a replacement
 for qmail-queue.
 
 The installation of a rather slow virus scanner on my own systems had lead
 me to realise a rare error condition I hadn't expected. This virus scanner
 didn't like scanning a 90Mb zip'ped AVI file (ahem) - whereas another vendor
 scanner took 1.5minutes to scan it, this one took nearly two hours...
 
 The sending SMTP server's qmail-remote timed out the SMTP session after 20
 minutes - as being in error - as it had waited "too long" for the final "OK".
 However, STDOUT on the receiving box still received the "mail from|rcpt to"
 envelope headers, so after 2 hours Qmail-Scanner happily delivered it back
 to the real qmail-queue for real delivery.

So let me get this right, what's happening is this:

o the remote site is connecting to qmail-smtpd

o qmail-smtpd is in turn invoking your replacement qmail-queue program
  called Qmail-Scanner

o Qmail-Scanner is in turn invoking the real qmail-queue.


Your problem arises when Qmail-Scanner (more correctly the scanner it
invokes I guess) takes a long time to process the data. In fact longer
than the SMTP timeout of the remote site. Then here's what happens:

o the remote site times out and closes the socket thinking the email
  delivery has failed

o meanwhile Qmail-Scanner et al are happily processing the email
  totally oblivious to the lost connection. Eventually the scan
  completes and the mail is injected into the local queue with
  qmail-queue.


The key is that Qmail-Scanner doesn't know that the socket has been
closed and that qmail-smtpd has exited.


My suggestion is that you take a two-pronged approach.

First off, introduce a timeout in Qmail-Scanner and exit accordingly
(exit(52) according to the qmail-queue man page).

Second off, I'd determine the process id of the parent with getppid()
and at the point at which the scan is complete - but just prior to
completing the qmail-queue - I'd use kill(parent, 0) to determine that
qmail-smtpd is still around.

All you are really doing is reducing the window of risk to a very
small - but non-zero - size. But non-zero is ok as SMTP is idempotent.


Your remaining problem is that the sender will never succeed as the
mail is too large to process within their SMTP time-frame, so a better
strategy might be to disconnect the scanner from SMTP. This is pretty
trivial with a two-instance qmail install but it sure adds complexity
for your customers.


Regards.



 
 However... back on the sending host, it tried to send it again...
 
 I had a little loop going there - quite nasty. Can you say "busy system"? :-)
 
 Anyhoo, the virus scanner is the real culprit here - and that's something
 that can be fixed (i.e. get another). The problem is WHY did the recipient
 qmail-smtpd send through the envelope headers via STDOUT to
 qmail-queue/Qmail-Scanner? Upon noticing the sender going away, shouldn't it
 have recognised that as an error condition?
 
 I'm gonna have to alarm Qmail-Scanner so it also spits the dummy before 20
 minutes (I hope other MTAs don't have shorter timeouts). That way it'll
 always be telling the sender MTA it's in trouble.
 
 Another solution would be to just accept the message before scanning it, and
 scan it after the sending server has gone away - but then I'd have to write
 an entire requeuing infrastructure to handle transient errors too (not
 bl**dy likely ;-)
 
 Oh yeah - and please don't say "limit the size" - we LIKE sending large
 things here :-) [we just don't appear to like receiving them ;-)]
 
 Am I missing something here? This seems to imply that if you had
 /var/qmail/queue on a VERY slow (but otherwise reliable) disk, that you
 would see this problem too. I hope I'm just been stupid and missed
 something obvious...
 
 
 -- 
 Cheers
 
 Jason Haar
 
 Unix/Special Projects, Trimble NZ
 Phone: +64 3 9635 377 Fax: +64 3 9635 417



Re: TCPServer Error

2001-04-17 Thread Mark Delany

On Tue, Apr 17, 2001 at 03:43:48AM -0300, Martin Marconcini wrote:
 Hello:
 
   I have followed www.lifewithqmail.org instructions. The server is OpenBSD 
 2.8. This was my first qmail installation. At the office I installed OpenBSD 
 and Qmail and followed instructions and have had no problem. I installed 
 pop/smtp stuff.
 
   At home I have another obsd box w/qmail. But I can't make tcpserver work.
 
   /var/log/qmail/smtpd/current shows the following error everytime I telnet 
 localhost 25.
 
 @40003ada6c2f381c5b64 tcpserver: status: 1/20
 @40003ada6c2f384a0ab4 tcpserver: pid 22092 from 127.0.0.1
 @40003ada6c2f38fae424 tcpserver: ok 22092 
 localhost.marconcini.com.ar:127.0.0.1:25 :127.0.0.1::17948
 @40003ada6c2f39080b54 tcpserver: warning: dropping connection, unable to 
 run /var/qmail/bin/qmail-smtpd: exec format error

The error message seems pretty obvious to me. There is some problem
with the qmail-smtpd executable. Perhaps it was compiled on a
different system, perhaps it was compiled on a later version of the
same system. Whatever the problem, your OS doesn't like that
executable file for some reason.

To confirm this, I'd run qmail-smtpd from a command line prompt
thusly:

$ /var/qmail/bin/qmail-smtpd

Please do this and show us the output.

To fix it you probably need to rebuild and reinstall the program. I
don't really know whether LWQ does this in the standard way.


 I have no inetd running. 

tcpserver is working fine. The problem is that the program it wants to
run (qmail-smtpd) is not running for some reason.


Regards.



Re: qmail-pop3d and supervise

2001-04-14 Thread Mark Delany

On Sat, Apr 14, 2001 at 01:38:03PM -0400, Rehan Zaidi wrote:
 Hi, folks.
 
 Thanks to the mailing list archives, I've been able to configure qmail-pop3d
 to run under supervise...almost.  I have one remaining problem: I still get
 "Connection refused" when I telnet to port 110.
 
 These are the processes running on the system:
 $ ps -ax | grep qmail
   160 ?S  0:00 supervise qmail-send
   163 ?S  0:29 supervise qmail-smtpd
   166 ?S  0:32 supervise qmail-pop3d
   167 ?S  0:00 qmail-send
   169 ?S  0:00 /usr/local/bin/multilog t s250
 /var/log/qmail/qma
   657 ?S  0:00 splogger qmail
   658 ?S  0:00 qmail-lspawn ./Mailbox
   659 ?S  0:00 qmail-rspawn
   660 ?S  0:00 qmail-clean

What is the output of:

$ ps -ax | grep tcp


 When I telnet to port 110, I get:
 telnet: Unable to connect to remote host: Connection refused

That tells us that tcpserver probably isn't running which means that
the run script is probably not running or runnable.

 But if I stop the qmail-pop3d and then start it from the command line using
 the following command, I can connect:
 /usr/bin/tcpserver -v -R 0 pop-3 /var/qmail/bin/qmail-popup myhost \
 /usr/local/bin/checkvpw /var/qmail/bin/qmail-pop3d Maildir
 
 This is the same thing as I have in the /var/qmail/supervise/qmail-pop3d/run
 script...

Is that script readable, executable?

What is the output of:

$ ls -l /var/qmail/supervise/qmail-pop3d/run

Also, rather then showing us the tcpserver command, much more
instructive would be a cat of the run file, so show us the output of:

$ cat /var/qmail/supervise/qmail-pop3d/run


Regards.



Re: Maildir file naming convention

2001-03-28 Thread Mark Delany

On Wed, Mar 28, 2001 at 09:52:58AM -0600, Charles Cazabon wrote:
 The right way to do it is clearly spelled out at:
 http://cr.yp.to/proto/maildir.html

That's true.

 It's necessary for all agents to use the same rules to prevent
 collisions.

Ok as far as it goes, but..

  The format is listed as follows,
  
  -rw---  1 subba  users 3599 Mar 28 07:32 985764747.20966_23.myhost:2,S
  -rw---  1 subba  users28883 Mar 28 01:55 __XE,5RUw6.myhost:2,S
 
 The first one is correct.  The second one does not follow djb's rules
 for naming the file.  If procmail wrote it, your version of procmail is
 broken.

I disagree. To quote from the webpage: "A unique name can be anything
that doesn't contain a colon (or slash) and doesn't start with a
dot.".

On that basis, the procmail filename is fine. Sure the webpage goes on
to *suggest* one method for generating unique names, but there is no
suggestion that that is the only way.

One could argue that procmail is being smart by ensuring that the
unique namespace it uses can only possibly collide with itself.

  How are these random names generated?

Anyway the MDA wants. The primary requirement is that it be
unique. You should not infer any meaning beyond uniqueness for
everything before the colon.

  Is this name generation the property of MUA such as mutt also? I
  thought it was the domain of MDAs.

Well, mutt lives within the rules of Maildir by only appending the
:info data to the filename rather than generating new filenames
(postponed messages notwithstanding).


Regards.



Re: faster than bcc

2001-03-28 Thread Mark Delany

On Wed, Mar 28, 2001 at 12:47:25PM -0800, Brett wrote:
 I remember reading that the fastest way to send one email to a large number
 of people is through bcc.

Well, the fact that it's Bcc: vs To: is not important wrt speed. The
reason for Bcc: over To: is to ensure that the recipient list isn't
visible to the recipients. That might have privacy implications and it
will certain have mail size implications with a million recipients!

 This was helpful to me because I'm not able to use
 a mailing list since the addresses I send to will be pulled dynamically from
 a database which is always changing. But somehow, populating the Bcc: field
 with a million names seems like it might not be the best idea to me. I
 understand qmail deletes this field before sending the message out but I'm
 more concerned with whether or not it will be making efficient use of the
 queue.

The performance gain comes from sending one mail with lots of
recipients. Those recipients traditionally are placed on Bcc: lines.

 Is the queue even used for one message sent to numerous people or is
 it only used for separate messages?

Both. The queue is *always* involved. However, one message with lots
of recipients creates much less work than lots of messages with one
recipient each - that's the key.

 If there's a better method than Bcc:-ing everyone, I'm very open to
 hearing it.

Not particularly. Some suggest usig qmail-queue directly (which
qmail-inject calls), but the interface is more difficult and the cost
saving is too small to measure for a large recipient list.

 One suggestion I got but which I
 can't get to work is:
 cat list.txt | xargs qmail-inject -a message.txt
 where list.txt is a list of addresses. Is this faster than Bcc: anyway? Any
 help much appreciated.

In what way can't you get it to work?

I would not use the xargs approach as that makes the recipients
visible and it is also less efficient than this:

( sed 's/^/Bcc: /' list.txt;cat message.txt ) | qmail-inject


Finally, make sure that message.txt has header lines, such as From:
and Subject: and make sure that there is an empty line between the
headers and the message text!


Regards.



Re: Delivered Messages staying in queue

2001-03-26 Thread Mark Delany

On Mon, Mar 26, 2001 at 03:49:15PM -0800, Bill Crowley wrote:
 Hi,
 
 It seems that delivered messages as staying in the queue. 7 days later queue
 mail is giving up "I'm not going to try again; this message has been in the
 queue too long."
 
 I know that 99% of these messages have been delivered successfully so I am
 not sure why they are not purging from the queue.

If it's a single email to multiple recipients, then qmail will not
delete the message until 100% of the messages have been delivered
successfully.

 Any help would be appreciated.

Show the list examples of what you means, especially useful is the
output from qmail-qread


Regards.



Re: Delivered Messages staying in queue

2001-03-26 Thread Mark Delany

On Mon, Mar 26, 2001 at 07:17:13PM -0500, Nick (Keith) Fish wrote:
 Peter van Dijk wrote:
  
  qmail-queue doesn't run as root. It runs as user qmailq. What group
  this user is in, or what his homedir is, doesn't matter. Permissions
  on the binary are relevant indeed.
  
  Greetz, Peter.
 
 Odd.  Why do I have a set-root-bit on my qmail-queue binary with an
 owner of qmailq, then?  I understand that it runs as user qmailq; but it
 runs with root's permissions, correct?  I have not modified it from the
 installation put in place by the tarball.

There is no such thing as a set-root-bit. There is a setuid bit, and a
setgid bit... their name deescribes their purpose.


Regards.



Re: redundant mail servers

2001-03-23 Thread Mark Delany

On Fri, Mar 23, 2001 at 08:09:01PM +0100, Vincent Schonau wrote:
 On Fri, Mar 23, 2001 at 11:01:51AM -0500, Russell Nelson wrote:
  Gopi Sundaram writes:
 
  I'm reluctant to move to Maildir until we can get more MUAs to support
  them (specifically Pine and Netscape).
 
  Wrong idea.  Never expose your mailboxes to your users.  Always use a
  virtual mailbox system -- either pop3 or imap.
 
 Why?

Flexibility. I have seem way too many mail systems start life as a
single box with a single disk grow into a multi-server setup.

Build in as much flexibility as you can from the start and you'll
never regret it. It costs nothing but a little thinking.

For example: if you only allow network access you can use load
balancing and DNS changes to move services around transparently. By
using network services you enable access by a much larger class of
client programs. By using network access you can transparently change
mailbox formats and server software.


Regards.




Re: Connection unexpectedly terminated

2001-03-22 Thread Mark Delany

On Thu, Mar 22, 2001 at 04:43:09PM -0600, Carey Jung wrote:
 Hi,
 
 We have a sporadic problem with qmail hanging and eventually timing out when
 popping certain messages.  tcpdump shows that qmail is apparently not
 handling the RETR command properly (see below).  Everything is fine until it
 "OK"'s the RETR command from the client, but then it immediately follows
 that with a FIN packet, terminating the connection.

Who sends the FIN? The client or the server?

 If we move the offending hung message out of the Maildir/new directory, then
 the client is able to pop the remaining messages.  If we put it back,
 send/receive stalls again.  We can not see anything out of the ordinary in
 the message itself, and, in fact, we are able to fetch it from other
 clients.

Sounds like a client bug. There are plenty of them? What's the client
OS, what's the client program?

I've certainly seen plenty of clients gag on unusual content.

Regards.


 
 Has anyone seen this before?  Why is qmail-pop3d/tcpserver terminating the
 connection without sending the mail to the client?
 
 tcpdump output
 --
 client  server: S 1646064:1646064(0)
 server  client: S 2361285140:2361285140(0) ack 1646065
 client  server: . 1:1(0) ack 1
 server  client: P 1:46(45) ack 1 (+OK
 [EMAIL PROTECTED])
 client  server: P 1:15(14) ack 46(user cmunson)
 server  client: . 46:46(0) ack 15(ack)
 server  client: P 46:52(6) ack 15(+OK )
 client  server: P 15:29(14) ack 52   (pass cm2ns0n)
 server  client: P 52:58(6) ack 29(+OK )
 client  server: P 29:35(6) ack 58(STAT)
 server  client: P 58:74(16) ack 35   (+OK 31 4646947)
 client  server: P 35:41(6) ack 74(UIDL)
 server  client: P 74:80(6) ack 41(+OK )
 server  client: P 80:616(536) ack 41 (UIDL data...)
 server  client: P 616:1073(457) ack 41 (more UIDL data...)
 client  server: . 41:41(0) ack 616   (ack)
 client  server: . 41:41(0) ack 1073  (ack)
 server  client: P 1073:1332(259) ack 41(remaining UIDL data...)
 client  server: P 41:50(9) ack 1332  (RETR 31)
 server  client: P 1332:1338(6) ack 50(+OK )
 server  client: F 1338:1338(0) ack 50  (close connection)
 client  server: . 50:50(0) ack 1339(ack)
 client  server: F 50:50(0) ack 1339(close connection)
 server  client: . 1339:1339(0) ack 51  (ack)
 
 environment:
 ---
 - qmail 1.0.3, w/tcpserver, vpopmail, etc.
 - Outlook 2000 client.  Also seen it once with Eudora.
 



Re: redundant mail servers

2001-03-21 Thread Mark Delany

 There will be 2 mail servers, mail1 and mail2
 
 Any email that is received by mail1 should automatically be forwarded
 to mail2, and any email that is received by mail2 should be forwarded
 to mail1. The only exception to the rule is when they receive messages
 from each other.
 
 Thus a user can check their email via IMAP or (shudder) POP from

Why shudder? POP is by far the most reliable service of the two and
much simpler and supported by more clients.

 either mail1 or mail2. If either server goes down, the other one
 should be receiving messages. The moment the server comes back up, it
 should receive all the messages that the other received during the
 down time.
 
 Is this a good way of providing redundancy? Or am I better off with a
 different mechanism?

This is not a very good mechanism particularly. First off, when they
delete an email on mail1, how will the copy on mail2 get deleted?

Second off, it seems that the user will have to know whether mail1 or
mail2 is the server that is available. That's not very user friendly.

 Can qmail be configured this way?

It can, but I doubt anyone will recommend such a setup.

The typical solution is to put the mailboxes of the users onto a
single, very reliable, piece of hardware (made reliable by redundancy
or high quality componentry or both), then use as many front-end
servers as needed to handle your redundancy requirements, load and
budget.

Remember, if the mailboxes are in Maildir format, they can safely be
shared across NFS. A simple configuration might be:

1.  A single high-availability NFS server - pick something that
supports RAID and has parts that can be replaced quickly and
easily.

This doesn't have to be something expensive like a Netapp -
though they are good for this. It could be something cheap
like an Intel BSD as long as you have spares on the
shelf. Don't use Linux for NFS serving - my experience is that
it's too buggy. Any of the other free Unixen will do the job -
pick the one you know best.

Spend as much money on this box as you can.


2.  A number of front end SMTP and POP servers. These front-end servers
mount the mailboxes from the NFS server. These front-end
servers don't need a lot of disk - just enough for the
mailq. Any of the free Unixen will do for this - pick the one
you know best.

3.  Use the DNS (or a load balancer if you have more money, but I note
the .edu address) to present these multiple front-end servers
as a single name/address to your user community. I recommend
something like smtp.yourdomain and pop.yourdomain.

 Since I have never set up qmail before, detailed explanations would be
 appreciated.

It's not really specific to qmail, but Maildir makes this a much more
viable solution compared to the locking and performance nightmares
associated with V7 mbox format used by sendmail and mail.local.


Regards.



Re: redundant mail servers

2001-03-21 Thread Mark Delany

On Wed, Mar 21, 2001 at 10:58:06AM -0500, Gopi Sundaram wrote:
 On 21 Mar 2001, Mark Delany wrote (quoting me):
 
   Thus a user can check their email via IMAP or (shudder) POP from
 
  Why shudder? POP is by far the most reliable service of the two
  and much simpler and supported by more clients.
 
 http://www.imap.org/papers/imap.vs.pop.brief.html

Right. My question remains. Why "shudder"? This article is 6 years old
and written by an IMAP proponent. Here's a couple of observations:

POP has turned out not be used mainly for "offline" mail processing.

The "offline vs online" model is largely dead these days.

Terry summarizes with: "its (IMAPs) additional complexity over POP
should not be a significant barrier to use."

I can't see how you shudder at POP on that basis. I agree that IMAP is
functionally richer, but that's about the only thing going for it.

  Remember, if the mailboxes are in Maildir format, they can safely
  be shared across NFS. A simple configuration might be:
 
 I'm reluctant to move to Maildir until we can get more MUAs to support
 them (specifically Pine and Netscape).

Are you talking about people who log into a shell or access via POP
and IMAP? If the latter, Maildir is transparent. If the former, you
never mentioned this, rather critical point.

 I've heard that the maildir format may have scalability issues because
 of the number of files that it deals with (bunches of open(), read()
 and stat() calls). Is there any truth to this?

This is tiresome FUD.

I can create a scenario that makes mbox look bad just as easily as a
scenario that makes Maildir look bad. Consider whether the FUD applies
to your scenario, not some imagined one created by a marketeer (and
yes geeks are just as guilty of marketing with FUD as the more
traditional salesdroid).


Regards.



Re: redundant mail servers

2001-03-21 Thread Mark Delany

On Wed, Mar 21, 2001 at 12:41:27PM -0500, Gopi Sundaram wrote:
 On Wed, 21 Mar 2001, Peter van Dijk wrote (quoting me):
 
   http://www.imap.org/papers/imap.vs.pop.brief.html
 
  And what is your *own* opinion? I prefer POP because IMAP makes
  users leave mail on server, amongst others.
 
 That is one of the reasons why I prefer IMAP. I don't like leaving my
 email lying on the various machines that I check my email from. There

POP does this too, if you choose.

 are several other reasons, but are irrelevent to this discussion,
 which follows:
 
  Uh. You are confused. Are you providing pop+imap or shell
  services?
 
 Both. And we have people that run Netscape on the mail server.

 That's what we have. What is *really big* ?  One of the aforementioned
 people had a 200MB mbox (which almost constantly crashed Netscape, and

It's not the size of the mailbox so much as the number of mails in the
mailbox. Depending on the file system, anything more than about
2-3,000 mails in a single mailbox will start to slow down a fair
amount.

 made Pine loop forever). I'm guessing that that won't be a problem if
 converted to maildir. I also read Mark Delany's post that dismisses my
 fears of scalability of the maildir format.

The point is that very few Maildirs reach the size where they fail
completely, on some file systems they just get very slow due to the
linear structure of the directory.

Do an experiment: run one of those mbox-to-maildir convert programs
(from www.qmail.org) on your 200MB mailbox - load it into a Maildir
and aim mutt at it and tell us what happens. Tell us how it compares
to pine loading it from mbox. Then delete one of the mails and
exit. Tells us how mutt performs and tell us how pine performs. In
fact do all your normal user interactions on each mailbox type and
share your results with us.

To do this experiment, all you need do is install mutt and download a
perl script. Surely a small price to pay to get some certainty for
yourself.

 Ideally, I would like mail to still be delivered to /var/mail/ in

Why do you want it in /var/mail particularly, apart from the fact that
you're used to it being there? If you're building a box from scratch
that is only a network service, I don't see where this requirement
comes from.

 whatever format, as long as I can get POP/IMAP servers to support it.
 Then users can read their email from NFS mounted spools when on our
 network, and via IMAP from anywhere else.

mbox is woeful across NFS. Try your 200MB mbox on an NFS server for a
while and draw your own conclusions. Remember that each open of an
mbox requires reading the whole mailbox and scanning from "From "
lines - all 200MB of it across the network. Opening a Maildir requires
reading the directory of the Maildir which is typically much smaller.

The idea of NFS mounting Maildir wasn't so that command line people
could get at it, it's so that other network service servers can share
it.

If people are using pine and netscape then can't both of these
programs be configured to acess a POP/IMAP server? In which case they
have no need to see the physical file structure. Once you move them
off the physical file structure onto a network service, you have
*much* greater flexibility.

 I guess if I use the maildir format, setting up redundant mailservers
 becomes easy. Here's my understanding:
 
 * equal priority MX records for two servers.
 * both servers running qmail, mail stored in an NFS mounted spool dir.
 * One or more servers that run IMAP/POP services that people can
   connect to (perhaps through one alias - mail.domain)
 
 Have I got it right?

Indeed.


Regards.



Re: handling bounces

2001-03-21 Thread Mark Delany

You can use VERP without using ezmlm. Checkout QMAILINJECT=r as
discussed in the qmail-inject manpage.

Regards.


On Wed, Mar 21, 2001 at 12:33:18PM -0800, Brett wrote:
 In qmail-inject, I'm Bcc-ing a LOT of people. What's the best method for
 handling bounces? I want to be able to extract a list of addresses into a
 file and deal with them later. I can't use ezmlm mainly because this Bcc
 list needs to be able to change on the fly (i.e. I can't just setup a static
 mailing list with ezmlm and have the bounce unsubscription automated through
 that since there's no such thing as a static mailing list in this
 situation). I've searched the usual places but can't find too much helpful
 info on this. Any help is appreciated, thanks.
 



Re: heavy traffic on port 25

2001-03-21 Thread Mark Delany

On Wed, Mar 21, 2001 at 11:30:26PM +0100, Krzysztof Wychowalek wrote:
 Dear friends,
 I have a server running Qmail as as MTA and about 300 mail 
 accounts. I realized that I'm experiencing huge amount of incoming 
 traffic to the port 25, it's like 1 MB per minute, so it slows down my 
 Internet connection dramatically. This is only incoming traffic, both 
 outgoing SMTP and POP3 is not more than 10-20 kB per minute. 
 But this big amount of data doesn't go to the users' mailboxes. It 
 goes... nowhere? I have no idea what it is actually. Even is 
 someone would use my server as an open relay, the amount of 
 incoming and outgoing SMTP packages would be more or less the 
 same.
 If someone has any idea, I would be very grateful for sending them 
 to me (priv). Thanks in advance.

What do the logs say? Mail doesn't just "go nowhere".


Regards.



Re: qmail large usuage

2001-03-20 Thread Mark Delany

On Tue, Mar 20, 2001 at 11:31:37PM +0100, Peter van Dijk wrote:
 On Wed, Mar 21, 2001 at 08:32:29AM +1100, [EMAIL PROTECTED] wrote:
  Hi, we do a mailout of about 40,000 - 50,000 emails per day to our
  clients and there clients (not spam). I have been trying to get qmail to
  work on getting up and over the 250 limitation of simulataneous
  connections.
  
  We are running hp netserver pIII 833 with 1 gig ram, the mail queue is
  running on raid 0. So I am sure we have the hardware to do it.
 
 I have a dual PIII-550 with 1 gig, queue on a dedicated scsi-disk. It
 sends out a mailinglist to the first 10.000 recipients in just over 3
 minutes, with concurrencyremote set to 256.
 
 So you should have no trouble at all  :)

My guess is that PeterM is sending unique emails, perhaps tailored TV
programs? My second guess is that PeterD is sending a single
untailored email to many recipients.


Regards.



Re: Control files

2001-03-19 Thread Mark Delany

On Mon, Mar 19, 2001 at 01:22:30PM -0800, Brad Dameron wrote:
 
 Is there a better description of what each file does in the
 /var/qmail/control directory?

Better than what exactly?

Better than "man qmail-control" which identifies all control files and
the relevant program in turn each have an individual manpage which
precisely descibes the use of each control file?


Regards.



Re: qmail-pop* and interface link

2001-03-15 Thread Mark Delany

On Thu, Mar 15, 2001 at 05:23:19PM +, Subba Rao wrote:
 I have qmail-popup and qmail-pop3d running on my system. Is it
 possible to dedicate this service to selected interfaces only?
 If it can be done, could you please point me to that URL?

Read up on tcpserver. That's the program that actually listens on the
interface. And yes, it can do it.


Regards.



Re: qmail reusing msg numbers - is this normal ?

2001-03-14 Thread Mark Delany

On Wed, Mar 14, 2001 at 01:14:43PM +, Greg Cope wrote:
 Dear All
 

 I.e msg no 325819 has been reused twice.
 
 Everything appears ok - is this something to worry about ?

No. It's entirely normal. The msg number is the inode. inodes get
reused by Unix when a file is deleted.


Regards.



Re: no local delivery???

2001-03-14 Thread Mark Delany

On Wed, Mar 14, 2001 at 10:27:14AM -0800, George Georgalis wrote:
 On Wed, Mar 14, 2001 at 01:38:07AM +, Mark Delany wrote:
 The output of qmail-showctl will tell you (and us) a lot more.
 
 that's what I like about qmail, so many programs to do just what you
 need! Just need to learn them now.
 
 
 I guess is that you have worldsite.ws in /var/qmail/control/me and
 something other than this domain in /var/qmail/control/locals
 
 
 You guessed right. Do I need to restart or HUP when I change these?

You're best bet is to read the qmail-control manpage which leads you
to which programs are affected by what control files and when/how they
notice changes.

 What's the best way to stop qmail? kill qmail-send?

Again, man qmail-send is your friend.


Regards.



Re: How to interpret the Delivered-To: header

2001-03-13 Thread Mark Delany

On Tue, Mar 13, 2001 at 08:46:19AM -0600, Charles Cazabon wrote:
 Norbert Bollow [EMAIL PROTECTED] wrote:
  How do you interpret the Delivered-To: header [...]
 [...] 
  Note: in the following, '**CENSORED**' replaces the localpart of

 If you want answers, don't hide the evidence that we may need to find them.
 You don't know the answer; that's why you're asking us -- so don't tell us
 what information we do and do not need to find your answer for you.

Indeed. I always find it amusing that people have a problem they
cannot solve, yet they know precisely what information is needed to
solve it.

Funny people.


Regards.



Re: How to interpret the Delivered-To: header

2001-03-13 Thread Mark Delany

 Well, there are very good reasons for avoiding to publicly
 post personal data about a subscriber to an infertility support
 group.

Fine. If it needs to remain confidential, buy support from someone
identified on www.qmail.org and have them sign an NDA. Problem solved.

If you pay for support - they abide by your terms. If you want free
support then we ask you to abide by our terms. That's not asking too
much is it?

 Anyway, if no one can answer the question based on the
 information which I have shared (which is very likely all the
 relevant data)

If you want to guess what we need then be my guest to guess away at
your solution, just don't ask us to guess.


Regards.




Re: no local delivery???

2001-03-13 Thread Mark Delany

The output of qmail-showctl will tell you (and us) a lot more.

I guess is that you have worldsite.ws in /var/qmail/control/me and
something other than this domain in /var/qmail/control/locals

Western Somoa huh? I had a lot of fun trying to register a domain
there, oh, 8 years ago.


Regards.


On Tue, Mar 13, 2001 at 05:22:14PM -0800, George Georgalis wrote:
 Hi,
 
 I'm installing qmail on LAN box and have not yet disabled sendmail.
 When I run either of these commands
 
  echo to: nonexistent | /var/qmail/bin/qmail-inject
  echo to: georgeg | /var/qmail/bin/qmail-inject
 
 mail goes through an external smtp. Why? I don't recall specifying the
 IP (or it's name) for any service... sendmail, uses different "smart"
 relay.
 
 Shouldn't qmail-inject drop into a local account?
 
 I've tried running ./config-fast with my box name (noe) and the domain
 I'll masq as (WorldSite.WS), with the same results.
 
 I'm also curious how WorldSite.WS got in the log; it's right but I don't
 recall specifying it...
 
 
 Mar 13 16:42:34 noe qmail: 984530554.069492 new msg 44712
 Mar 13 16:42:34 noe qmail: 984530554.069728 info msg 44712: bytes 203 from 
[EMAIL PROTECTED] qp 3743 uid 500
 Mar 13 16:42:34 noe qmail: 984530554.075204 starting delivery 7: msg 44712 to remote 
[EMAIL PROTECTED]
 Mar 13 16:42:34 noe qmail: 984530554.075384 status: local 0/10 remote 1/20
 Mar 13 16:42:34 noe qmail: 984530554.544747 delivery 7: success: 
216.35.187.251_accepted_message./Remote_host_said:_250_2.0.0_f2E0gXd13122_Message_accepted_for_delivery/
 Mar 13 16:42:34 noe qmail: 984530554.544936 status: local 0/10 remote 0/20
 Mar 13 16:42:34 noe qmail: 984530554.544989 end msg 44712
 
 Thanks!
 // George
 
 -- 
 George Georgalis, System Administrator http://WorldSite.WS
 Global Domains International 701 Palomar Airport Rd, Suite 300
 Carlsbad, CA 92009 U.S.A.  Phone: 760.602.3000 Fax: 760.602.3099



Re: qmail-pop3d bug

2001-03-12 Thread Mark Delany

  A more sensible strategy might be to introduce a new "info" flag (say
  '3' equals POP wire size) on the filename, eg, a 10,000 byte email has
  a name something like this:
  
  Maildir/new/980195114.16740.geex:2,RS3,1
 
 From reading URL:http://cr.yp.to/proto/maildir.html, it is not clear to me
 that this would be the proper format for such an 'info' extension. I would
 worry that MUAs and other software dealing with maildir (scripts!) would
 expect info semantics in the 2, series to be at the end of the filenames.

Indeed, and given that "info is morally equivalent to the Status field
used by mbox readers" I suspect that the my suggested syntax is beyond
the original intent.

  Optimally the wire-size is calculated when the mail is written to
  Maildir/tmp/ and then applied as an "info" flag when the file is moved
  to Maildir/new/.
 
  A possible complication with this approach is that my reading of
  Maildir infers that "info" can only be set when the file moves from
  Maildir/new/ to Maildir/cur/.
  
 No, this is not what that document says. It says
 
"When you move a file from new to cur, you have to change it's name [...]"

You stopped quoting before the most important part! Here's the
complete sentence.

"When you move a file from new to cur, you have to change its name
from uniq to uniq:info."

To me that implies that a file in new cannot have an "info" section.

 You *have* to change the name when the file move from new/ to cur/ , but
 there is no specification of other cases; in fact, lots of MUA's will change
 info when the file has been in cur/ for a while: mutt, for example, moves
 the file from new/ to cur/, adds :2, and only modifies that to be 2,S after
 the user has read the message (it is no longer 'N'ew).

Right, but that's my point. To specify another case.


Regards.




Re: Please help!!!

2001-03-12 Thread Mark Delany

On Mon, Mar 12, 2001 at 01:52:00AM -0800, Sean Coyle wrote:
 Also, 
 
 Another good thing to note:
 
 I was having a serious problem with qMail delivering mail to end users,
 however, mail was being stored in the queue (this was quite some time ago
 now).  I had made a few changes to a crontab entry the day earlier, however;
 everything was seemingly running normally with qMail.
 
 Anyway, it turns out that I was running an INTENCE CRON job every single
 second,

Are you sure about that? All the crons I've seen only let you run a
job at most, once per minute. How did you get cron to run something
once per second?

If the job runs at most once per minute, it's hard to imagine how it
would consume all available resources such that qmail stopped
delivering.

(Of course it's not impossible, but just unlikely: a job that runs for
an hour that is started once per minute may well have a serious
resource impact).

 Just an example of how a completely unrelated system event can alter the
 performance of other items... (hrrmmm chaos mathematics anyone?)

Most likely the problem is much simpler and more directly related. As
the earlier poster suggested a systematic process of elimination is
the best approach.


Regards.



Re: qmail-pop3d bug

2001-03-11 Thread Mark Delany

On Sat, Mar 10, 2001 at 01:12:13PM -0500, John R Levine wrote:
 The usual mailbox vs. maildir war has flared up on inet-access, and points
 out a bug in qmail-pop3d.  When you do a LIST command, it gives you the
 size of each message.  Pop3d just reports the file sizes, while it's clear
 from the RFC that it's supposed to report the wire size of each message,
 i.e., the size using cr/lf as a line terminator, so the sizes it reports
 are too small.
 
 I gather nobody's ever reported this as a bug, and I expect that the only
 thing that uses the size is the "don't download bigger than size X" option
 for which it's close enough, but it's still wrong.

If I mis-remember correctly, qpopper may have a similar problem in
that the stated size does not necessarily match the size sent down the
wire. How so?  Because qpopper adds X-UIDL and Status: headers to the
out-going message (perhaps it includes this in the size calc but I
haven't looked at the code in such a long time, or perhaps it only
adds those headers when the mail is re-written).

 I use courier-imap, and its POP daemon does get the sizes right,
 presumably by reading the files and adding the number of \n characters.

A more sensible strategy might be to introduce a new "info" flag (say
'3' equals POP wire size) on the filename, eg, a 10,000 byte email has
a name something like this:

Maildir/new/980195114.16740.geex:2,RS3,1

Optimally the wire-size is calculated when the mail is written to
Maildir/tmp/ and then applied as an "info" flag when the file is moved
to Maildir/new/.

A possible complication with this approach is that my reading of
Maildir infers that "info" can only be set when the file moves from
Maildir/new/ to Maildir/cur/.


Regards.



Re: qmail-pop3d bug

2001-03-11 Thread Mark Delany

 Yes. This behaviour is known. Fixing it, however, involves a *huge*
 performance downgrade of qmail-pop3d.

Not if it's calculated as the file is written to the Maildir.

 'Usually, during the AUTHORIZATION state of the POP3 session, the POP3
 server can calculate the size of each message in octets when it opens
 the maildrop. . simply counts each occurance of this character in
 a message as two octets.'

Typical of those RFCs authors that, consciously or otherwise, used a
single implementation to guide much of their thinking on protocol
design. POP3 is not the only standard that suffers as a consequence -
consider SMTP and DNS?

We shouldn't have to live with short-sightedness forever.


Regards.



Re: qmail-pop3d bug

2001-03-11 Thread Mark Delany

On Sun, Mar 11, 2001 at 06:05:47PM -0500, John R. Levine wrote:
 Putting the linecount in there makes more sense. Some MUAs might be happy
 about that, and it still allows easy calculation of wiresize (add
 number of lines to physical size). More info, less bytes :)
 
  Optimally the wire-size is calculated when the mail is written to
  Maildir/tmp/ and then applied as an "info" flag when the file is moved
  to Maildir/new/.
 
 Yes. Mind the performance penalty tho.
 
 Not a bad idea.

Agreed. Line count is probably a more useful number as the other
values can be derived. I retract my POPsize suggestion in favour of
line count.

 The performance penalty would be tiny, reading buffers
 that are about to be written out won't cause an extra page fault.

I also agree that it's an acceptable CP cost to scan a buffer just
prior to writing. CP is cheap and plentiful on most qmail systems.

  A possible complication with this approach is that my reading of
  Maildir infers that "info" can only be set when the file moves from
  Maildir/new/ to Maildir/cur/.
 
 That's what the spec says, indeed. A delivery process is not supposed
 to know anything, so :info is not needed in new/.
 
 Gee, we find that even Dan isn't infallible.  In retrospect, there's all
 sorts of hints that the delivery process could leave.

Yep. And it probably wouldn't be too hard to change the standard
though I note that, eg, mutt totally ignores any existing "info"
values. But I'm willing to bet that they will change code if they see
a good reason and they will be especially interested in a change that
lets them know line count without scanning.


Regards.



Re: [Fwd: Administrivia: Mailing List Software]

2001-03-10 Thread Mark Delany

On Sat, Mar 10, 2001 at 11:49:08PM +0100, Andre Oppermann wrote:
 Hey guys,
 
 lets make this poor man happy and let us all tell him about how well
 qmail/ezmlm works!
 
 This guy is Elias Levy (aleph1) and he runs the Bugtraq mailing list.
 
 Please send an email directly to him if you want to suggest qmail/ezmlm
 for running a large mailing list with a secure piece of software. And
 he also is sick of handling bounces...

Whilst bounce processing is indeed a sale point for ezmlm, much of
what Elias wants is above and beyond ezmlm. For example categorization
and subscription by category. Sure you can (painfully) make a sublist
for each category, as long as they don't invent and rename categories
on the fly.

Elias also talks about an emulation layer for LISTSERV. I've not heard
of anyone providing that for ezmlm.

This is not to under-rate ezmlm, as a base toolkit it would perform
admirably, but the BUGTRAQ dood wants a lot of value-adds that are not
part of ezmlm.


Regards.



Re: pop3d needs SUID root?

2001-03-09 Thread Mark Delany

On Fri, Mar 09, 2001 at 11:26:58PM +, Ricardo Cerqueira wrote:
 On Fri, Mar 09, 2001 at 03:06:08PM -0800, Todd A. Jacobs wrote:
  When running pop3d, I get an error saying "no $HOME/Maildir" which may be
  caused by the directory permissions of 0700 on both the home directory and
  Maildir.
  
  On Red Hat, procmail is the MDA, and is SUID/SGID root. Other than making
  pop3d run as root, what are my options? If I chmod the directories, what's
 
 qmail-pop3d IS supposed to run as root. From LWQ:

Nope.

 tcpserver -v -R 0 pop3 /var/qmail/bin/qmail-popup FQDN \
 /bin/checkpassword /var/qmail/bin/qmail-pop3d Maildir 21 | \
 /var/qmail/bin/splogger pop3d 

tcpserver runs as root, qmail-popup inherits root, checkpassword
inherits root but changes to the uid/gid of the user that successfully
logged in, pop3d inherits the uid/gid that checkpassword changed to.


Getting back to the very confusing question. What has procmail
setuidness got to do with pop3d?

On the permissions front you forgot to mention the owner of
$HOME/Maildir. Is it owned by the user? It should be.

Finally, if you have procmail delivering to the users $HOME/Maildir
then it does not need to be setuid root.


Regards.



Re: pop3d needs SUID root?

2001-03-09 Thread Mark Delany

On Fri, Mar 09, 2001 at 03:33:17PM -0800, Aaron L. Meehan wrote:
 Quoting Todd A. Jacobs ([EMAIL PROTECTED]):
  When running pop3d, I get an error saying "no $HOME/Maildir" which may be
  caused by the directory permissions of 0700 on both the home directory and
  Maildir.
 
 qmail-pop3d is run as root by tcpserver, which is running as root.  No
 suid bit is needed.  qmail-pop3d switches to the userid of the user
 whose mail it is retrieving.

Nope. checkpassword does the switch, qmail-pop3d runs as whatever user
it inherits.

 So, you'll need to start believing that message--qmail-pop3d can't
 find the user's Maildir.

Wise words indeed.

 Your tcpserver's command line may be goofed up.  Should look
 something like:

Also check:

1.  Does the user have a $HOME/Maildir
2.  Does the user have access to this dir (could be owned by root)


Regards.



Re: running qmail from /supervise

2001-03-05 Thread Mark Delany

 is svscan.  For example, here's my /service directory on my server:
 
 axfrdns   dnscache  ftpd  msql2dqmail rsyncdsshd
 bray  etrn  httpd pop3d qmtpd smtpd tinydns
 
 Most of these are obvious.  "bray" is not a service name but instead
 the name of a friend who needs to have a tcpservice running on a
 non-root port.  So I do this:
 #!/bin/sh
 exec setuidgid bray   tcpserver  -HRl0 0 2379 ~bray/bin/server

Nothing wrong with this, but just as an alternative thought, nothing
stops users running their own svscan. I have a ~/service and guess how
the per-user svscans are started...


Regards.



Re: Qmail and time zone

2001-03-04 Thread Mark Delany

On Sun, Mar 04, 2001 at 09:43:28AM -0500, Kari Suomela wrote:
 
 Sunday March 04 2001 05:36, Mark Delany wrote to Kari Suomela:
 
 
  MD As others have said, qmail only puts a Date: header in if one 
  MD isn't
  MD already present,
 
 That's probably what it should be doing, except it's not doing it 
 right.

According to which particular standard?

 The Date header should include the TZ, i.e. GMT offset.

According to which particular standard?

Btw. Personal preference does not count as a standard.


Regards.



Re: New qmail version request

2001-03-03 Thread Mark Delany

On Fri, Mar 02, 2001 at 10:14:20PM -0600, Charles Cazabon wrote:
 Mark Delany [EMAIL PROTECTED] wrote:
   
   But why does qmail have to be patched to use LDAP?  Why not use a script
   which extracts user information from the LDAP database, puts it in passwd
   format, and feeds it to qmail-pw2u?  Then cron it every hour or something.
   Voila,
  
  Better yet, why not make a replacement qmail-getpw? That's how I built an
  LDAP-aware qmail a couple of years ago.
 
 But if the LDAP query fails in qmail-getpw-ldap, you have to either defer or
 bounce.

Sure. But there's nothing wrong with a deferral.

If deferrals are a problem for qmail-getpw-ldap, I'd pursue a more
reliable LDAP service. Remember that the LDAP service is most likely
also used to authenticate your POP users and they'll what
authentication to be reliable well before qmail cares about a few
deferrals.

In any event what I was really getting at was merely the modularity
that a qmail-getpw plugin can provide.


Regards.



Re: Qmail and time zone

2001-03-03 Thread Mark Delany

On Sat, Mar 03, 2001 at 11:28:30PM -0500, Kari Suomela wrote:
 
 Thursday March 01 2001 22:41, David Dyer-Bennet wrote to All:
 
   No, it's not! That's how I noticed it. Someone was blaming my 
   client
   for it, but the problem is the same with all of them. I have tested
   it with various Netscapes, Outlook 98, Outlook 2000, Outlook
   Express, PMMail Pro 2000, Sqwebmail and Adjewebmail.
 
  DB That's because you didn't use a client which adjusts header
  DB timestamps, though.
 
 I am not talking about clients! Mail generated on a qmail server 
 doesn't have proper date headers, whereas mail coming from a sendmail 
 server does.

Er, what do you mean by "proper date headers" and how are you sure you
definition of "proper date headers" isn't being met by qmail?

I suspect what is happening is that qmail is creating Date: headers
that are UTC based and you are used to seeing Date: headers in your
local time zone. Are you sure that what qmail is doing is incorrect or
is it's possible that it's legal according to the standards, but just
that it's different from what you want?

If it's legal according to the standards, but differs from what you
expect, what's your problem exactly?

As others have said, qmail only puts a Date: header in if one isn't
already present, so you can easily override the default by using a
program that puts in a Date: field.

Regards.



Re: New qmail version request

2001-03-02 Thread Mark Delany

On Fri, Mar 02, 2001 at 05:17:01PM -0600, Charles Cazabon wrote:
 Chris Garrigues [EMAIL PROTECTED] wrote:
  
   Much of the common patches that are around fail in one of the tests above,
   at least when using the author's stringent tests.  There's nothing wrong
   with this; he keeps qmail secure, reliable, efficient, and "correct", and
   anyone who wants to applies patches as they see fit.
  
  I, for one, am hoping that 2.0 will have LDAP support which meets his
  standards.  
 
 As you said, the existing LDAP libraries are probably crap.  But why does
 qmail have to be patched to use LDAP?  Why not use a script which extracts
 user information from the LDAP database, puts it in passwd format, and
 feeds it to qmail-pw2u?  Then cron it every hour or something.  Voila,

Better yet, why not make a replacement qmail-getpw? That's how I built
an LDAP-aware qmail a couple of years ago.

One problem with replacing qmail-getpw is that the domain isn't know.
which is a problem for multi-domain systems, so I modified
qmail-lspawn to pass the domain to qmail-getpw. The code is no big
deal, but I'm hopeful DJB will consider the idea in a future release
as it increases the ease with which alternative user databases can be
supported in an unmodified qmail.


Regards.



Re: Lost the Battle

2001-03-01 Thread Mark Delany

On Thu, Mar 01, 2001 at 10:19:34AM -0500, Dave Sill wrote:
 My qmail project, only 1 week away from implementation, was canned, we are
 now moving to Lotus Notes.
 
 Well, it's not a total loss. At least you learned something about
 qmail.

And maybe you can convince your company to use qmail as your email
relay server on the firewall. Use Notes internally in a protected
environment and only expose qmail to that nasty world out there.

Sure you could expose your Notes server to the Internet, but do you
really want to with all that company data so close at hand?

Sure you could also buy a seperate Notes server and license just as a
firewall box, but is that cost effective and is it the most secure
choice?


Regards.



Re: qmail-conf-054 / multilog

2001-02-23 Thread Mark Delany

On Fri, Feb 23, 2001 at 12:13:37PM -0500, Paul Farber wrote:
 I have
 
 drwxr-xr-x   4 root root 4096 Feb 21 01:40 service
 
 and under that
 
 drwxr-xr-x   4 root root 4096 Feb 21 01:40 .
 drwxr-xr-x  13 root qmail4096 Feb 20 23:20 ..
 drwxr-sr-x   5 root root 4096 Feb 21 01:11 qmail
 drwxr-sr-x   5 root root 4096 Feb 21 01:35 qmail-smtpd
 
 Since the manual (man chmod) or the qmail-conf program docs didn't allude
 to it... what is the 'sticky bit'?

Hmm. Solaris manpage talks about it as does FreeBSD - you must be on
Linux, right? In octal it is 1000 or symbolically, +t. Thus:

chmod +t /service/qmail


Regards.


 
  
 
 Paul Farber
 Farber Technology
 [EMAIL PROTECTED]
 Ph  570-628-5303
 Fax 570-628-5545
 
 On Fri, 23 Feb 2001, Chris Johnson wrote:
 
  On Fri, Feb 23, 2001 at 10:33:57AM -0500, Paul Farber wrote:
   It appears that all logging is being dumped to the first virtual console:
   
   info msg 224981: bytes 32957 from #@[] qp 9474 uid 0
   starting delivery 1692: msg 224981 to local [EMAIL PROTECTED]
   status: local 1/10 remote 0/20
   delivery 1692: success: did_1+0+0/
   status: local 0/10 remote 0/20
   end msg 224981
   new msg 224981
   info msg 224981: bytes 3235 from
   [EMAIL PROTECTED] qp 9479 uid 0
   starting delivery 1693: msg 224981 to local [EMAIL PROTECTED]
   status: local 1/10 remote 0/20
   delivery 1693: success: did_1+0+0/
   status: local 0/10 remote 0/20
   end msg 224981
   
   even though I have log/run set up EXACTLY as in the supervise man page.
   Is this a supervise/multilog bug?  anyone getting qmail to log deliveries
   to multilog using log/run ?
  
  Did you set the sticky bit on your qmail service directory?
  
  Chris
  
 



Re: cyclog line?

2001-02-23 Thread Mark Delany

On Fri, Feb 23, 2001 at 01:02:57PM -0600, Charles Cazabon wrote:
 Bill Parker [EMAIL PROTECTED] wrote:
  
  Now according to the man page for cyclog:
  
  cyclog [ -ssize ] [ -nnum ] [ -mmargin ] dir
 [...] 
  Is there a space needed between the -s or not?
 
 No.  You don't believe TFM?

And... What happens when you try it both ways? Such an experiment
won't create world hunger - give it a try and report and discrepancies
with the documentation back to the list.

Sometimes a simple experiment is going to give you a more reliable
learning experience than just asking a list.


Regards.



Re: qmail-conf-054 / multilog

2001-02-23 Thread Mark Delany

Assuminmg you're running this all via svscan, the problem is that
svscan only notices the +t flag when it first sees the directory in
/service.

You need to remove the service and re-add it. I believe the
daemontools page at cr.yp.to has the sequence needed to do this.


Regards.


On Fri, Feb 23, 2001 at 03:05:23PM -0500, Paul Farber wrote:
 Tried that...
 
 
 but it will not fire off a copy of multilog:
 
  3181 ?S  0:00 supervise qmail-smtpd
  3182 ?S  0:00 /usr/local/bin/tcpserver -qDHR 
 -ladmin.f-tech.net -xt 
 
 here is some file info:
 
 drwxr-sr-t   5 root root 4096 Feb 21 01:35 qmail-smtpd
 
 ./run=
 
 #!/bin/sh
 exec 21 \
 envdir ./env \
 sh -c '
 case "$REMOTENAME" in h) H=;; p) H=p;; *) H=H;; esac
 case "$REMOTEINFO" in r) R=;; [0-9]*) R="t$REMOTEINFO";; *) R=R;; esac
 exec \
 /usr/local/bin/envuidgid qmaild \
 softlimit ${DATALIMIT+"-d$DATALIMIT"} \
 /usr/local/bin/tcpserver \
 -qD"$H$R" \
 ${LOCALNAME+"-l$LOCALNAME"} \
 ${BACKLOG+"-b$BACKLOG"} \
 ${CONCURRENCY+"-c$CONCURRENCY"} \
 -xtcp.cdb \
 -- "${IP-0}" "${PORT-25}" \
 /var/qmail/bin/qmail-smtpd
 '
 
 ./run/log=
 
 [root@admin log]# cat run
 #!/bin/sh
 exec \
 setuidgid qmaill \
 multilog t ./main
 
 ./mail is empty... and there is no multilog process running
 
 Paul Farber
 Farber Technology
 [EMAIL PROTECTED]
 Ph  570-628-5303
 Fax 570-628-5545
 
 On Fri, 23 Feb 2001, Chris Johnson wrote:
 
  On Fri, Feb 23, 2001 at 12:13:37PM -0500, Paul Farber wrote:
   drwxr-xr-x   4 root root 4096 Feb 21 01:40 service
   
   and under that
   
   drwxr-xr-x   4 root root 4096 Feb 21 01:40 .
   drwxr-xr-x  13 root qmail4096 Feb 20 23:20 ..
   drwxr-sr-x   5 root root 4096 Feb 21 01:11 qmail
   drwxr-sr-x   5 root root 4096 Feb 21 01:35 qmail-smtpd
   
   Since the manual (man chmod) or the qmail-conf program docs didn't allude
   to it... what is the 'sticky bit'?
  
  chmod +t the directory. (Are you sure man chmod doesn't refer to this?)
  
  From http://cr.yp.to/daemontools/svscan.html:
  
  "If a subdirectory sub is sticky, svscan starts a pair of supervise processes,
  one for sub, one for sub/log, with a pipe between them. svscan needs two free
  descriptors for each pipe."
  
  Chris
  
 



Re: Inter7 introduces new software: vQregister

2001-02-23 Thread Mark Delany

On Fri, Feb 23, 2001 at 11:30:53AM -0600, [EMAIL PROTECTED] wrote:
 Maybe what I said wasn't as clear as it could
 have been.  Exactly what you requested below,
 is the feature we will be adding.
 
 PHP is inefficient BTW. :)

Totally OT, but one user registration per second adds up to 86,400 new
users per day.

Can a small web server running PHP handle one registration per second?
Answer: yes.  Does hotmail.com do more than 86K registrations per day?
Answer: no.

Conclusion: One small webserver running PHP can handle all the
registrations for arguable the largest webmail service on the planet.

Efficiency is not always the most relevant selection criteria.


Regards.

 
 Dan Phoenix wrote:
  
  Quite honestly this is a custom form that most of us code ourselves in php
  to insert the info we need from users. What you really need to do is allow
  us to pick our own fields...and integrate whatever html we want into it.
  
  Regards,
  
  Dan
  
  On Fri, 23 Feb 2001 [EMAIL PROTECTED] wrote:
  
   Date: Fri, 23 Feb 2001 09:52:06 -0600
   From: [EMAIL PROTECTED]
   To: [EMAIL PROTECTED], [EMAIL PROTECTED], [EMAIL PROTECTED]
   Subject: Inter7 introduces new software: vQregister
  
   vQregister is new web-based signup CGI that
   more than replaces the old vQsignup program
   we released early last year.  If you're running
   vQsignup now, or you're considering allowing
   users to signup for free accounts, it's worth
   taking a look at.
  
   We will be adding a new feature very soon
   which will be used to collect demographic,
   or any other information administrators might
   want during the signup process.  This information
   might include their cleartext password, a
   challenge password (for retrieving the cleartext
   password over the phone), and other demographic
   information such as zip codes, cities, etc.
   We intend to make this fully configurable, so
   that whatever information you wish to collect,
   is easy to setup.
  
   Head over to http://www.inter7.com/vqregister to
   take a look at it.
  
   --
   [EMAIL PROTECTED]
   Inter7 Internet Technologies, Inc.
   www.inter7.com - 847-492-0470
   New prices!  http://www.inter7.com/prices.html
  
 
 -- 
 [EMAIL PROTECTED]
 Inter7 Internet Technologies, Inc.
 www.inter7.com - 847-492-0470
 New prices!  http://www.inter7.com/prices.html



Re: qmail-conf-054 / multilog

2001-02-23 Thread Mark Delany

For debugging purposes you might want to run svscan manually so the
errors go to the screen/window you're on.


On Fri, Feb 23, 2001 at 06:25:11PM -0500, Paul Farber wrote:
 Is there a specific kernel setting I need for supervise to log

Not unless it's some wierd Unix. svscan writes errors to stderr. You
don't need kernel settings to control where that goes.

 correctly??? I am on kernel 2.4.1 it's working fine on a RH 6.2
 machine with 2.2.17 (using djbdns).
 
 The sticky bit seemed to have no effect..

Did you restart svscan?

 and nothing is being logged error-wise.

Did you check the system console?


Regards.

 
 Paul Farber
 Farber Technology
 [EMAIL PROTECTED]
 Ph  570-628-5303
 Fax 570-628-5545
 
 On Fri, 23 Feb 2001, Chris Johnson wrote:
 
  On Fri, Feb 23, 2001 at 12:13:37PM -0500, Paul Farber wrote:
   drwxr-xr-x   4 root root 4096 Feb 21 01:40 service
   
   and under that
   
   drwxr-xr-x   4 root root 4096 Feb 21 01:40 .
   drwxr-xr-x  13 root qmail4096 Feb 20 23:20 ..
   drwxr-sr-x   5 root root 4096 Feb 21 01:11 qmail
   drwxr-sr-x   5 root root 4096 Feb 21 01:35 qmail-smtpd
   
   Since the manual (man chmod) or the qmail-conf program docs didn't allude
   to it... what is the 'sticky bit'?
  
  chmod +t the directory. (Are you sure man chmod doesn't refer to this?)
  
  From http://cr.yp.to/daemontools/svscan.html:
  
  "If a subdirectory sub is sticky, svscan starts a pair of supervise processes,
  one for sub, one for sub/log, with a pipe between them. svscan needs two free
  descriptors for each pipe."
  
  Chris
  
 



Re: Concurrency questions

2001-02-18 Thread Mark Delany

  In this way you'll make the first delivery attempt yourself for each
  recipient; avoiding any overhead in the qmail-send process or the queue
  management. if the first attempt fails then the message is passed off to
  qmail-send to handle, which should be a much lower volume of mail.
 
 I understand this code has to be executed in a loop for each
 recipient...Can you explain what advantages I get doing it this way ?

By using qmail-remote directly you avoid all of the I/O overhead of
placing the message in the queue and have qmail-send find and process
each message. Going to qmail-remote directly is a zero I/O cost
strategy.

I recently did a system somewhat like this across multiple servers and
the systems doing the initial qmail-remote delivery attempt where
diskless. Works a treat.


Regards.



Re: Concurrency questions

2001-02-18 Thread Mark Delany

 pass the message off to qmail to deliver. As most message get delivered on
 the first attempt you'll save the overhead of writing the message to disk,

And this is a large caveat. If, eg, your network happens to be down at
the time you attempt delivery, you'll inject a huge number of emails
into qmail - that may hurt. When I've developed this sort of code I've
found it just as easy to do the retries in the qmail-remote driver
logic and dispense with qmail-inject/qmail-send altogether. Risk
avoidance is the motive.

 duplicates and other things like that, but still it's much less work
 than the absolutle requirements for reliable delivery a general
 purpose mail delivery agent has.

As always, Richard knows what he speaks about. There are many
optimizations available when you have very specific requirements that
are less demanding than a general purpose mail delivery system such as
qmail. The really great news is that you can use the qmail componentry
such as qmail-remote to reduce your development costs. Try that with
sendmail!


Regards.



Re: Syslog? [was Re: Detail logging of POP3D]

2001-02-12 Thread Mark Delany

On Mon, Feb 12, 2001 at 12:17:47PM -0500, Paul Farber wrote:
 djb has several logging options... lately I believe it is multilog(?) in
 the new daemontools package.

Indeed. And it's pretty triv to use too. The simplest is to replace
your 'splogger qmai' with something like:

multilog t n10 s500 /var/log/qmail

Make sure you have a /var/log/qmail directory and that's about it.

Of course if you want to go the whole hog and use svscan/supervise,
then it's a little different as you create a log/run script etc, but
the above is the simple way.

   We've heard this again and again. Any specifics?
  
  I've seen the syslog daemon simply die.  With no explanation.  Several times
  on different boxes.  I think this qualifies as being unreliable.

Indeed syslogd notoriously dies on Solaris 2.5 and 2.6 especially when
hit with multiple HUPqs at around the same time (which of course
happens quite a lot if you use multiple log rolling scripts as one is
almost forced to do on Solaris with that /usr/lib/newsyslog
abomination).


Regards.



Re: multi-thread

2001-02-08 Thread Mark Delany

On Thu, Feb 08, 2001 at 12:06:05PM -, Tim Goodwin wrote:
  ok, on my Solaris, the qmail distribution is "forking" almost 10 to 20
  processes per second.
  
  This cost a lot in system ressources and system calls
 
 Yes.  Unfortunately, Solaris isn't Unix, and qmail was designed to run
 on Unix systems.  Unix is rather good at forking, especially images as
 tiny as qmail; Solaris isn't.  As Rob Pike once said, "perhaps if
 people had understood fork() better we wouldn't have threads".
 
  So I'm trying to work on a threaded qmail-rspawn to avoid so many forks
 
 Yikes.

If all he's trying to achive is reduce forking on his Solaris box, I
concur. However if we generalize the question, I don't know that I'd
draw the same conclusion.

If any area of qmail would benefit for threading, it might be the
remote delivery mechanism - currently handled by Batman and Robin, er,
sorry, qmail-rspawn and qmail-remote.

First off, there is an amount of data they can share and cache, such
as tcpok and recent DNS lookups.

Second, remote delivery can have very high latency so any footprint
saving is a big saving.

Third, the state requirements are truly tiny. A socket and an fd is
just about all that the thread needs.

Fourth, there are few security issues. Neither qmail-rspawn nor
qmail-remote need any special file system access. This is often a
nasty complication for threaded implementations. Not so here.

Fifth, the interface is simple and clean, plug in the threaded
qmail-rspawn and no one is any the wiser.

Sixth, the problem domain isn't that large:

$ wc -l qmail-rspawn.c qmail-remote.c

 103 qmail-rspawn.c
 427 qmail-remote.c
 530 total

Having said that, in the scheme of things, qmail-remote borders on
ridiculously tiny as it is. I recently wrote a queueless wrapper
program that uses qmail-remote as the smtp engine (opt-in spam I call
it). I rediscovered that a concurrency of 1,000 qmail-remotes consumes
very little system resource on FreeBSD.

 I'm going to put my manager's hat on for a moment.  How much time do
 you intend to spend on developing and debugging this?  How much does
 that time cost?  How much would it cost to buy a fast PC, run a real
 Unix (I'd suggest OpenBSD, FreeBSD, or some version of Linux) on it,
 and make that your mail server?

As a solitary exercise solely designed to speed up one system, of
course replacing the box may be a better solution.


Regards.



Re: multi-thread

2001-02-08 Thread Mark Delany

  Fifth, the interface is simple and clean, plug in the threaded
  qmail-rspawn and no one is any the wiser.
 
 With nonblocking sockets and select(), one could write a
 single-threaded qmail-rspawn/remote. Only need to find a way to do the
 dns lookups in parallel.

Yes Virginia. There are at least three ways to skin a cat in Unix. And
even the design of a good threaded implementation may not be the
obvious one of one thread per socket. It might be, eg, a thread per
function, or a thread per lockable area or a thread per external
interface or...

And yes Virginia. This is off topic now.


Regards.




Re: compiling qmail-1.03 under SCO Open Server 5.05 and the UDK from SCO.

2001-02-07 Thread Mark Delany

On Wed, Feb 07, 2001 at 04:08:38PM +, Uwe Ohse wrote:
 On Wed, Feb 07, 2001 at 09:17:10AM -0500, Jocelyn Clement wrote:
  
  This is it: I ran the "make setup check" and it generates an error
  message on the "qmail-local.c" saying that there is no definition
  of the "timestruct_t" in the "stat.h" file.
  
   I am using the SCO development system.
  
   Please share with us ALL the information you could possibly have on
  hand on  "How to" compile qmail under SCO Open Server 5.05
 
 well, i'm actually only cursed with one last 5.0.2 system without 
 any working development system, but ...

I'm with Uwe on this front. My experiences with qmail/tcpserver on SCO
(sorry don't know the exact version now) were anything but fun. When
you finally compile it, you'll want to search the archives for SCO - I
vaguely recall posting something on this regarding the need for
tcpserver -o.


Regards.



Re: Perl checkpassword

2001-02-06 Thread Mark Delany

On Tue, Feb 06, 2001 at 12:17:44PM -0600, Larry M. Smith is the BPFH wrote:
 Someone had asked for this some time ago... But I forget who or when.
 
 DJB, if you would, please archive locally to www.qmail.org.

That would be [EMAIL PROTECTED], aka Russ Nelson, but I'm sure he'll
see the message.


Regards.



Re: no incoming mail from outside

2001-02-06 Thread Mark Delany

Well, usually the log files will tell you what's going on... what do
they say?


 any ideas? Clean Redhat 7.0 install with all the latest qmail and vpopmail.
 I have followed the instructions to the letter 4 times and I can't seem to

 cat rc
 #!/bin/sh
 
 # Using stdout for logging
 # Using control/defaultdelivery from qmail-local to deliver messages by
 default
 
 # These following are the defaults from LWQ mail - doesn't work with
 vpopmail
 
 [ross@ws1 qmail]$ cat rc
 exec env - PATH="/var/qmail/bin:$PATH" \
 qmail-start .Mailbox

Default delivery is to .Mailbox? Is that what the instructions really
say? Are you over-riding this with per-user .qmail files? If not, you
have a mismatch between your delivery type and the pop server as:


 exec env - PATH="/var/qmail/bin:/usr/local/bin" \
 tcpserver -H -R 0 pop-3 \
 /var/qmail/bin/qmail-popup mail.innsandcottages.com \
 /home/vpopmail/bin/vchkpw /var/qmail/bin/qmail-pop3d Maildir 

The qmail pop server only reads Maildirs - it knows nothing about
.Mailbox

You may find that your logs are showing delivery, but you're maybe
expecting the mail to be delivered into a Maildir and it's not?


Regards.



Re: High MEM Usage??

2001-02-04 Thread Mark Delany

Well, this is hardly a qmail question. It's more a system
administration/Linux question. Have you got the 'top' command? Try
that? Have you got the 'ps' command? Try that.

I don't know about Linux so much, but some Operating Systems use
memory that has never had anything placed in it in preference to
memory that has had something loaded into it.

What that means is that if you run 100 different programs, rather than
reuse the one piece of memory for each program, the OS will load in
the first program, leave it in memoryt, and load in the next program
at the next available piece of memory. Over time this has the effect
of using all your memory, but of course the OS is just being smart
about caching. That may be all that's happened with your system.

Relating to qmail. qmail is a very small consumer of memory and is
unlikely to be relevant to any interpretation you are making on this
output.


Regards.


On Sun, Feb 04, 2001 at 08:04:47PM +0530, Sumith Ail wrote:
 Hello,
 
 We have just received our server which is a Dual PIII with 512 MB RAM , RH Linux 6.2 
Box. I have installed qmail on this with tcpserver, Now the meminfo shows
 cat /proc/meminfo
 
 total:used:free:  shared: buffers:  cached:
 Mem:  529530880 364380160 165150720 72847360 300982272 24657920
 Swap: 10485514240 1048551424
 MemTotal:517120 kB
 MemFree: 161280 kB
 MemShared:71140 kB
 Buffers: 293928 kB
 Cached:   24080 kB
 BigTotal: 0 kB
 BigFree:  0 kB
 SwapTotal:  1023976 kB
 SwapFree:   1023976 kB  
 
 There is hardly anybody using this server...please let me know how can I find out 
which process is using so much of memory.
 
 Kind Regards
 Sumith
 



Re: High Mem usage??

2001-02-04 Thread Mark Delany

Er, one copy of this email to the list is more than enough. Three is
clearly excessive.


Regards.


On Sun, Feb 04, 2001 at 06:31:44AM -0800, Sumith Ail wrote:
 Hello,
 
 We have just received our server which is a Dual PIII with 512 MB RAM , RH Linux 6.2 
Box. I have installed qmail on this with tcpserver, Now the meminfo shows
 cat /proc/meminfo
 
 total:used:free:  shared: buffers:  cached:
 Mem:  529530880 364380160 165150720 72847360 300982272 24657920
 Swap: 10485514240 1048551424
 MemTotal:517120 kB
 MemFree: 161280 kB
 MemShared:71140 kB
 Buffers: 293928 kB
 Cached:   24080 kB
 BigTotal: 0 kB
 BigFree:  0 kB
 SwapTotal:  1023976 kB
 SwapFree:   1023976 kB  
 
 There is hardly anybody using this server...please let me know how can I find out 
which process is using so much of memory.
 
 Kind Regards
 Sumith
 



  1   2   3   4   >