Specific configuration of sendmail to ignore 5.*.* codes
dryuk
I have very specific task and need help to realize it.
My server(solaris) is running sendmail(8.11) for sending some auto-
generated messages to another mail server(this server is not on my
control). Receiving server sometimes get troubles and sending e-mail
get answer 5.*.*(for example User Unknown). Atfter that e-mail does
not try to deliver again and generate ndr. And we lose this mail.
Can I modify the behavior of sendmail for 5.*.*( not generate ndr but
try to send message again and again)?
Andrzej Adam Filip
Have you considered sending the messages directly or via your custom
script without help of local MTA/sendmail?
What kind of program/script sends the messages?
How much "configurability" do you have?
--
[pl>en Andrew] Andrzej Adam Filip : an...@onet.eu : Andrze...@gmail.com
Open-Sendmail: http://open-sendmail.sourceforge.net/
Criticism comes easier than craftsmanship.
-- Zeuxis
dryuk
> Have you considered sending the messages directly or via your custom
> script without help of local MTA/sendmail?
>
> What kind of program/script sends the messages?
> How much "configurability" do you have?
> Open-Sendmail:http://open-sendmail.sourceforge.net/
> Criticism comes easier than craftsmanship.
> -- Zeuxis
Thank you for your answer.
This message are generated from other servers and send to my as "smart
host".
Current scheme of sendmail configuration
1rst process works -bd option and mode=queue.He recieves messages and
using "queugroup" option queued the to two diffferent spool by domains
(test.ru - /var/spool/mqueue/test, other - /var/spool/mqueue).Other
parameters of configuration is standard.
2nd process works -q5m option in the default mode. He takes message
from /var/spool/mqueue/test and send this messages to "problem host".
Specific options is Privacy_Flags=nobodyreturn,noreciept, and modified
timeouts of message spooling from 5 days to 20
3rd process works -q5m option in the default mode. He takes message
from /var/spool/mqueu and works with these messages.
I am interested in changing configuration of 2nd process.
Andrzej Adam Filip
1) About how many messages "per peak hour" are you talking about?
[Some "scripting" solutions I think about would be unfit for 100+]
2) Have you considered using "pull" instead of "push" deliveries?
e.g. IMAP server with IDLE support on your "smart host"
[ IDLE to get "real time" deliveries ]
--
[pl>en Andrew] Andrzej Adam Filip : an...@onet.eu : Andrze...@gmail.com
Open-Sendmail: http://open-sendmail.sourceforge.net/
Coincidences are spiritual puns.
-- G. K. Chesterton
dryuk
> [Some "scripting" solutions I think about would be unfit for 100+]
>
> 2) Have you considered using "pull" instead of "push" deliveries?
> e.g. IMAP server with IDLE support on your "smart host"
> [ IDLE to get "real time" deliveries ]
1. About 200 messages.
Thank you for ideas of workaround, but I want to try solving this
problem using only sendmail.
If it imposssible, use workaround.
Have any ideas about modifying sendmail configuration?
Andrzej Adam Filip
AFAIK making sendmail treat 5?? replies in a non standard way would
require patching sendmail sources and can not be done via sendmail.cf.
What you want is too far away from "mainstream" sendmail use.
--
[pl>en Andrew] Andrzej Adam Filip : an...@onet.eu : Andrze...@gmail.com
Open-Sendmail: http://open-sendmail.sourceforge.net/
Be wary of strong drink. It can make you shoot at tax collectors and miss.
-- Lazarus Long, "Time Enough for Love"
dryuk
> dryuk <dryu...@gmail.com> wrote:
> >> 1) About how many messages "per peak hour" are you talking about?
> >> [Some "scripting" solutions I think about would be unfit for 100+]
>
> >> 2) Have you considered using "pull" instead of "push" deliveries?
> >> e.g. IMAP server with IDLE support on your "smart host"
> >> [ IDLE to get "real time" deliveries ]
>
> > 1. About 200 messages.
>
> > Thank you for ideas of workaround, but I want to try solving this
> > problem using only sendmail.
> > If it imposssible, use workaround.
> > Have any ideas about modifying sendmail configuration?
>
> AFAIK making sendmail treat 5?? replies in a non standard way would
> require patching sendmail sources and can not be done via sendmail.cf.
>
> What you want is too far away from "mainstream" sendmail use.
>
Grant Taylor
> Can I modify the behavior of sendmail for 5.*.*( not generate ndr but
> try to send message again and again)?
I'm sure it's possible to hack Sendmail to do this.
What you are wanting to do not only goes against Sendmail (or any other
normal MTA for that matter) but it also goes against SMTP and how they
are designed.
/If/ you do this I would strongly suggest that you NOT do this on your
production mail servers. Rather set up a special mail server / daemon
that has been modified this way and route mail to the receiving server
in question via this modified server. Doing this will allow the rest of
your email to flow as intended while still sending the problem emails to
an instance of Sendmail that does what you want it to do.
Grant. . . .
D. Stussy
news:5ff1c43a-8d4f-480d...@z7g2000vbh.googlegroups.com...
Read the proposal for the "ExtendedErrorDrop" option posted 2 weeks ago.
ska
Hm, are you trying to mail bomb a server?
When this remote server gets in trouble, how many connections does
your local box has opened?
I would first try to limit the number of connections to the server to
"1" (one).
For instance by queue-only your mails (sendmail -odq) and make sure
there is running one queue runner at a time.
Bye, ska