Hello...
I fixed some quirks with my previous rpm and I am releasing another
version.
The following are at ftp://admin.netus.com *see note at bottom.
sendmail-8.9.1a-5.i386.rpm
sendmail-8.9.1a-5.src.rpm
sendmail-8.9.1a.spec
sendmail-cf-8.9.1a-5.i386.rpm
sendmail-doc-8.9.1a-5.i386.rpm
Intel only at the moment but it should compile cleanly on both alpha and
sparc.
These contain the 'a' patch from
ftp://ftp.sendmail.org/pub/sendmail/sendmail.8.9.1a.patch. They also
contain the patch sendmail_linux.diff. And mime.c.patch, patch.orbs,
proto.m4.p1, and proto.m4.p2 from
http://www.informatik.uni-kiel.de/%7Eca/email/patches . ORBS and RBL
are cool!
The sendmail.cf (from the .mc) file contains
FEATURE(access_db, hash -o /etc/mail/access)dnl
FEATURE(always_add_domain)dnl
FEATURE(blacklist_recipients)dnl
FEATURE(limited_masquerade)dnl
FEATURE(local_procmail)dnl
FEATURE(masquerade_entire_domain)dnl
FEATURE(orbs)dnl
FEATURE(rbl)dnl
FEATURE(redirect)dnl
FEATURE(relay_based_on_MX)dnl
FEATURE(relay_entire_domain)dnl
FEATURE(relay_local_from)dnl
FEATURE(use_ct_file)dnl
FEATURE(use_cw_file)dnl
MAILER(smtpdemanddial)dnl
MAILER(procmail)dnl
MAILER(smtp)dnl
MAILER(usenet)dnl
MAILER(uucp)dnl
relay_local_from is not as good as pop-before-smtp but that will come
later.
smtpdemanddial is for dialin servers. Originally from Harker Systems,
but I changed it some. Now the way to use it is from the mailer table (
dialin_donainname.com qsmtp:dialin_domainname.com).
If you customers that do this you need a way to queue mail from all
three queues when they dial in. I use the script etrn_domainname
(available via ftp) run from radiusd (also available via ftp).
The basic of the whole system is:
Run one sendmail daemon to accept and process incoming connections.
Run another daemon to process the queue in /var/spool/mqueue at 10
minute intervals.
Run re-mqueue (from sendmail contrib) to move older messages to another
queue.
Move messages older than 1 hour from mqueue to mqueue2 and messages in
mqueue2 older than 4 hours to mqueue3.
Run another daemon to process the queue in /var/spool/mqueue2 at 65
minute intervals.
Run another daemon to process the queue in /var/spool/mqueue3 at 4 hour
intervals.
I am running this unaltered on my secondary machine. It runs very
well. I am running it without ORBS and RBL on my primary machine. I
will put ORBS and RBL in as soon as the DNS agreements come back. :)
Is anyone interested in a linux High-throughput sendmail HOWTO? (at
least the way I do it)
*note at bottom :)
I designed these mostly for an ISP mail server (ours specifically).
They are probably overkill for a home/low usage machine.
These rpms put almost everything in /etc/mail (see _CF_DIR_/README).
They will _NOT_ upgrade cleanly. You need to make copies of your
existing sendmail configs just in case. Most (All) of you existing
configs will be replaced. If you have made changes to them they should
be re-named {file}.rpmsave or {file}.rpmorig. You need to move these to
the new location and start sendmail. (you did stop it before upgrading,
yes). You should try this on a test machine first. Again: !You should
try this on a test machine first!
Feedback is welcome. (positive and negative, but positive would be
better)
--Christopher McCrory Lead Bithead, Netus Inc. chrismcc@netus.com admin@netus.com
"Linux: Because rebooting is for adding new hardware"