🚀 go-pugleaf

This is accomplised through the sendmail.conf file. I would strongly advise
checking on one of O'Riely books before editing it though.


Thanks
Aaron N. Martin

Andrew Taylor <news@hot-shot.demon.co.uk> wrote in message
news:933521112.11557.0.nnrp-10.c2dea8e2@news.demon.co.uk...
> Hi,
>
> I have a 99% working sendmail config on a Mandrake 6 box. What makes
> Sendmail queue it's external mail until my ppp connection is up but
deliver
> internal mail imediatley?
>
> Andy
>
>
>

In article <933521112.11557.0.nnrp-10.c2dea8e2@news.demon.co.uk>, 
news@hot-shot.demon.co.uk says...

> I have a 99% working sendmail config on a Mandrake 6 box. What makes
> Sendmail queue it's external mail until my ppp connection is up but deliver
> internal mail imediatley?

I have define(`confDELIVERY_MODE',`b') set in my .mc file. This means 
that delivery is attempted in the background. With my setup, local 
mail is delivered immediately. For other mail, if I am online it gets 
tossed over to post.d.c.u, otherwise it gets queued automatically due 
to DNS lookup failures. In ip-up.local I do a sendmail -q that sends 
the mail when the PPP link comes up.

Other values are

define(`confDELIVERY_MODE',`q') - queue always
define(`confDELIVERY_MODE',`d') - deferred delivery
define(`confDELIVERY_MODE',`i') - immediate

The O'Reilly Bat book is excellent and explains all of these things, 
although it is not for the faint hearted. 

Also checkout www.sendmail.org for the FAQ and tips

Anjan

>>>>> "Andrew" == Andrew Taylor <news@hot-shot.demon.co.uk> wrote:

   Andrew> I have a 99% working sendmail config on a Mandrake 6
   Andrew> box. What makes Sendmail queue it's external mail until my
   Andrew> ppp connection is up but deliver internal mail imediatley?

The default setup AFAIK.  If the ppp connection isn't up sendmail
should simply queue external mail anyway because it can't look them up
in DNS or connect to the host.

The only thing you might want to do is ensure sendmail isn't running
the queue every 15 minutes, or whatever, by making sure it starts as
just `sendmail -bd' rather than `sendmail -bd -q15', and then just run
the queue when you connect with `sendmail -q' in /etc/ppp/ip-up.

Stefan

On Sun, 1 Aug 1999 16:27:21 +0100, Andrew Taylor <news@hot-shot.demon.co.uk> 
wrote:

>I have a 99% working sendmail config on a Mandrake 6 box. What makes
>Sendmail queue it's external mail until my ppp connection is up but deliver
>internal mail imediatley?

In /etc/sendmail.cf have this line:

# default delivery mode
O DeliveryMode=background

and make a file /etc/service.switch containing the lines:

hosts	files dns
aliases files


This works on RedHat 5.2 with sendmail upgraded to 8.9.1 which will I
think be pretty similar to what you're using.

-- 
Simon Child       Spam trap in Email Address.
Send email to simon.at.child.demon.co.uk  (Replace .at.)
Send spam to postmaster@127.0.0.1

In demon.ip.support.unix stefan the exiled <smd@hopf.demon.co.uk> wrote:
>>>>>> "Andrew" == Andrew Taylor <news@hot-shot.demon.co.uk> wrote:

>    Andrew> I have a 99% working sendmail config on a Mandrake 6
>    Andrew> box. What makes Sendmail queue it's external mail until my
>    Andrew> ppp connection is up but deliver internal mail imediatley?

> The default setup AFAIK.  If the ppp connection isn't up sendmail
> should simply queue external mail anyway because it can't look them up
> in DNS or connect to the host.

> The only thing you might want to do is ensure sendmail isn't running
> the queue every 15 minutes, or whatever, by making sure it starts as
> just `sendmail -bd' rather than `sendmail -bd -q15', and then just run
> the queue when you connect with `sendmail -q' in /etc/ppp/ip-up.

I've got a question here. Two questions, in fact. One a sendmail
question, one a suck question.

I configured my current set-up using the very helpful guides
on www.hopf.demon.co.uk, and it works rather well, but for two
things.

At the moment, I'm using kppp to initiate a ppp connection. When
I initially connect, none of the mail can get through, because 
I've got sendmail set to check the mail domains resolve properly,
but the punts seem to attempt delivery before access to demon's
dns servers kicks in. So sendmail just keeps refusing to deliver,
offering up continuous 'cannot resolve' messages in the mail log.

I kill the sendmail demon, and then re-launch it, and everything
is fine. So, my question is, is there some way I can stop the
punts from attempting to deliver immediately, or does anyone have
any other clues what I should do to fix this?

Question number two, is my suck question. For some reason,
I can't seem to run suck from within a script. I've tried to
call it from withing ip-up, I've tried it from a script of 
it's own. Nada. I'm probably missing something completely
obvious, but I've no idea what it might be. It runs fine if
I just type the same command by hand.

Clues please?

Peter McDermott <peter@petermc.demon.co.uk> wrote:
> I've got a question here. Two questions, in fact. One a sendmail
> question, one a suck question.
> 
> I configured my current set-up using the very helpful guides
> on www.hopf.demon.co.uk, and it works rather well, but for two
> things.
> 
> At the moment, I'm using kppp to initiate a ppp connection. When
> I initially connect, none of the mail can get through, because
> I've got sendmail set to check the mail domains resolve properly,
> but the punts seem to attempt delivery before access to demon's
> dns servers kicks in. So sendmail just keeps refusing to deliver,
> offering up continuous 'cannot resolve' messages in the mail log.
> 
> I kill the sendmail demon, and then re-launch it, and everything
> is fine. So, my question is, is there some way I can stop the
> punts from attempting to deliver immediately, or does anyone have
> any other clues what I should do to fix this?

You could get sendmail to not check the domains, add to .mc file:
FEATURE(accept_unresolvable_domains)

Another way might be to have your own cacheing DNS and use ipchains
to control access to it, i.e. add/remove rules in ip-down/ip-up.


> 
> Question number two, is my suck question. For some reason,
> I can't seem to run suck from within a script. I've tried to
> call it from withing ip-up, I've tried it from a script of
> it's own. Nada. I'm probably missing something completely
> obvious, but I've no idea what it might be. It runs fine if
> I just type the same command by hand.

Is it requiring an environment variable to be set which isn't
set when called from ip-up?

Peter McDermott wrote:
> 
> In demon.ip.support.unix stefan the exiled <smd@hopf.demon.co.uk> wrote:
> >>>>>> "Andrew" == Andrew Taylor <news@hot-shot.demon.co.uk> wrote:
> 
> >    Andrew> I have a 99% working sendmail config on a Mandrake 6
> >    Andrew> box. What makes Sendmail queue it's external mail until my
> >    Andrew> ppp connection is up but deliver internal mail imediatley?
> 
> > The default setup AFAIK.  If the ppp connection isn't up sendmail
> > should simply queue external mail anyway because it can't look them up
> > in DNS or connect to the host.
> 
> > The only thing you might want to do is ensure sendmail isn't running
> > the queue every 15 minutes, or whatever, by making sure it starts as
> > just `sendmail -bd' rather than `sendmail -bd -q15', and then just run
> > the queue when you connect with `sendmail -q' in /etc/ppp/ip-up.
> 
> I've got a question here. Two questions, in fact. One a sendmail
> question, one a suck question.
> 
> I configured my current set-up using the very helpful guides
> on www.hopf.demon.co.uk, and it works rather well, but for two
> things.
> 
> At the moment, I'm using kppp to initiate a ppp connection. When
> I initially connect, none of the mail can get through, because
> I've got sendmail set to check the mail domains resolve properly,
> but the punts seem to attempt delivery before access to demon's
> dns servers kicks in. So sendmail just keeps refusing to deliver,
> offering up continuous 'cannot resolve' messages in the mail log.
> 
> I kill the sendmail demon, and then re-launch it, and everything
> is fine. So, my question is, is there some way I can stop the
> punts from attempting to deliver immediately, or does anyone have
> any other clues what I should do to fix this?
> 

The the tip at http://www.harker.com/sendmail/smtpdemanddial.html

You may have to tweak for 8.9.x but the idea is sound.


> Question number two, is my suck question. For some reason,
> I can't seem to run suck from within a script. I've tried to
> call it from withing ip-up, I've tried it from a script of
> it's own. Nada. I'm probably missing something completely
> obvious, but I've no idea what it might be. It runs fine if
> I just type the same command by hand.
> 
> Clues please?

I think you've been caught by insufficient configuration. The following
section of the FAQ should get you on track:
http://www.sendmail.org/faq/section3.html#3.11
-- 
________________________________________________________________________________
G. Roderick Singleton, <gerry@pathtech.org> PATH tech,
71 Underhill Drive, Unit 159, Toronto, ON  M3A 2J8
Voice : 416-452-4583 Fax: 416-452-0036
________________________________________________________________________________

In demon.ip.support.unix Paul Black <paul@canix.co.uk> wrote:
> Peter McDermott <peter@petermc.demon.co.uk> wrote:
>> 
>> I kill the sendmail demon, and then re-launch it, and everything
>> is fine. So, my question is, is there some way I can stop the
>> punts from attempting to deliver immediately, or does anyone have
>> any other clues what I should do to fix this?

> You could get sendmail to not check the domains, add to .mc file:
> FEATURE(accept_unresolvable_domains)

I'd considered this, though I wanted to avoid it as I
seem to be getting quite a lot of spam recently and
this appears to bounce a fair amount of it. 


> Another way might be to have your own cacheing DNS and use ipchains
> to control access to it, i.e. add/remove rules in ip-down/ip-up.

Ah. I actually am running my own caching DNS (though
I haven't gotten around to setting up ipchains yet.
Could it be that what's happening is that it tries
to resolve the waiting mails before a connection to
demon's dns is established?

Is it essential to firewall my dns? It's not a service
I would have thought likely to be accessed by outside
domains?
 
>> 
>> Question number two, is my suck question. For some reason,
>> I can't seem to run suck from within a script. I've tried to
>> call it from withing ip-up, I've tried it from a script of
>> it's own. Nada. I'm probably missing something completely
>> obvious, but I've no idea what it might be. It runs fine if
>> I just type the same command by hand.

> Is it requiring an environment variable to be set which isn't
> set when called from ip-up?

I don't believe so. I get exactly the same results when I
run the command in a shell script as root, yet it works 
perfectly well if I type it in.

Those enviroment variables are a serious pain in the arse
though. They've caught me out umpteen times in the past.
I'll read the man page again, see if I'm missing anything
obvious.

Peter McDermott <peter@petermc.demon.co.uk> wrote:
> Is it essential to firewall my dns? It's not a service
> I would have thought likely to be accessed by outside
> domains?

The firewall isn't to protect DNS, it's to prevent access by your
internal machine/network. The reason I do it is to speed things up
when I'm not connected to the Internet. If DNS is running all the
time, some applications pause while a DNS lookup times out. By
using the following two commands:
/sbin/ipchains -D input -j REJECT -p udp -d gammon domain
/sbin/ipchains -I input -j REJECT -p udp -d gammon domain
I effectively turn off DNS (it's the same as stopping then
starting, only quicker) and it preserves some cached DNS data which
also speeds things up later on (gammon is the gateway machine).

Paul

[comp.mail.sendmail removed from followups]

Peter McDermott wrote:
> 
> In demon.ip.support.unix Paul Black <paul@canix.co.uk> wrote:
> > Peter McDermott <peter@petermc.demon.co.uk> wrote:
> >>
> >> I kill the sendmail demon, and then re-launch it, and everything
> >> is fine. So, my question is, is there some way I can stop the
> >> punts from attempting to deliver immediately, or does anyone have
> >> any other clues what I should do to fix this?
> 
> > You could get sendmail to not check the domains, add to .mc file:
> > FEATURE(accept_unresolvable_domains)
> 
> I'd considered this, though I wanted to avoid it as I
> seem to be getting quite a lot of spam recently and
> this appears to bounce a fair amount of it.
> 
> > Another way might be to have your own cacheing DNS and use ipchains
> > to control access to it, i.e. add/remove rules in ip-down/ip-up.
> 
> Ah. I actually am running my own caching DNS (though
> I haven't gotten around to setting up ipchains yet.
> Could it be that what's happening is that it tries
> to resolve the waiting mails before a connection to
> demon's dns is established?
> 
> Is it essential to firewall my dns? It's not a service
> I would have thought likely to be accessed by outside
> domains?
> 
> >>
> >> Question number two, is my suck question. For some reason,
> >> I can't seem to run suck from within a script. I've tried to
> >> call it from withing ip-up, I've tried it from a script of
> >> it's own. Nada. I'm probably missing something completely
> >> obvious, but I've no idea what it might be. It runs fine if
> >> I just type the same command by hand.
> 
> > Is it requiring an environment variable to be set which isn't
> > set when called from ip-up?
> 
> I don't believe so. I get exactly the same results when I
> run the command in a shell script as root, yet it works
> perfectly well if I type it in.
> 
> Those enviroment variables are a serious pain in the arse
> though. They've caught me out umpteen times in the past.
> I'll read the man page again, see if I'm missing anything
> obvious.


In reading Peter's comments, I remembered that in SysV and its
comtemporaries shells behave differently.  I.e. setuid stuff ends up
not setuid. For Sun's ksh the following is the flag to use:

          -p   Disables processing of the $HOME/.profile file and
               uses the file /etc/suid_profile instead of the ENV
               file.  This mode is on whenever the effective  uid
               is  not  equal to the real uid, or when the effec-
               tive gid is not equal to the  real  gid.   Turning
               this  off  causes  the effective uid and gid to be
               set to the real uid and gid.

Just a shot in the dark.
-- 
________________________________________________________________________________
G. Roderick Singleton, <gerry@pathtech.org> PATH tech,
71 Underhill Drive, Unit 159, Toronto, ON  M3A 2J8
Voice : 416-452-4583 Fax: 416-452-0036
________________________________________________________________________________

In demon.ip.support.unix G. Roderick Singleton <gerry@pathtech.org> wrote:
> Peter McDermott wrote:
>> 
>> Those enviroment variables are a serious pain in the arse
>> though. They've caught me out umpteen times in the past.
>> I'll read the man page again, see if I'm missing anything
>> obvious.


> In reading Peter's comments, I remembered that in SysV and its
> comtemporaries shells behave differently.  I.e. setuid stuff ends up
> not setuid. For Sun's ksh the following is the flag to use:

Well, I finally worked it out. Just another of those dumb
mistakes that I make all the time.

I was starting my scripts with #!/usr/bin/sh and only
have bash on this machine. Every other distro I've used
has had a symbolic link from sh to bash, but not this
one, apparently. :-(

Thanks for the help though.