The 03.05.03 at 14:19, Patrick wrote:
think. Thus, you could increase those delays, as long as total delay is below 30". Maybe 4" and 3" (RETRANS/RETRY) would be safer :-?
The question is, how long one try is during. If it is - and that's what I assume - 10 seconds, then a RETRY of 3 is already too much to stay below 30s. I don't know, if the 10s are right, but if I do a manual nslookup to a 'dead' RBL-server, it waits for about 10s until I get a "No response from server".
Then, trial and error for you, I'm afraid.
Thanks for the hint! The setup of syslog was no problem so far. But I cannot verify, if the "-v" parameter was passed to sendmail.
In "/etc/sysconfig/sendmail" I added a "-v" to the ARGS variable
SENDMAIL_ARGS="-v -L sendmail -Am -bd -q30m -om"
and restartet sendmail; but this didn't work. I also tried to add "-v" directly to "/etc/init.d/sendmail" with
if test -z "$SENDMAIL_ARGS" ; then SENDMAIL_ARGS="-v -L sendmail -Am -bd -q30m -om"
but this didn't help either.
Funny... Just stop the sendmail service - assuming you can, of course - and start up sendmail manually on a console. There you should be able to easily try out the options. I can not check them, because I have postfix installed [hold on] I just booted my old computer, with suse 7.3 and sendmail. -v Go into verbose mode. Alias expansions will be announced, etc. -L tag Set the identifier used in syslog messages to the supplied tag. **> -A not found! -bd Run as a daemon. This requires Berkeley IPC. Sendmail will fork and run in background listening on socket 25 for incoming SMTP connections. This is normally run from /etc/rc. -q[time] Processed saved messages in the queue at given intervals. If time is omitted, process the queue once. Time is given as a tagged number, with `s' being seconds, `m' being minutes, `h' being hours, `d' being days, and `w' being weeks. For example, `-q1h30m' or `-q90m' would both set the timeout to one hour thirty minutes. If time is specified, sendmail will run in the background. This option can be used safely with -bd. -ox value Set option x to the specified value. This form uses single character names only. The short names are not described in this manual page; see the Sendmail Installation and Operation Guide for details. So, -om is unknown. Also, you could use: -X logfile Log all traffic in and out of mailers in the indi cated log file. This should only be used as a last resort for debugging mailer bugs. It will log a lot of data very quickly. Maybe verbose is not an option in daemon mode - after all, I used it when I called it directly from the ip-script to send pending mail. In any case, case, syslog debug info should work anyway: if a file for debug info is not defined, it just is not written, but discarded. In your case, to get the extra info you need, you might have to use the -X option in debug mode.
I still have the same messages in "/var/log/mail.debug" as I had (before changeing syslog.conf) in "/var/log/mail". I also tried "-d", "-d8.1", "-d99.100 -d8.1" but nothing special showed up in "/var/log/mail.debug". Only with the "-d" parameter some debug-output appeared on the console when starting sendmail with "rcsendmail start".
I also tried "-X /var/log/mail.traffic" which gave me exactly what I am looking for! But unfotunately without any timestamp. So I'm still unable to verify the time-behaveiour of sendmail during name lookups... What did I wrong?
Ah! so you tried it. No time stamp? Mmm... you could use cron to add a line every five minutes to that file, with a time stamp. Kind of a hack, but it might work. Argh, no, you need to discriminate seconds... Ideas... :-? Ask at a sendmail forum, list, wahtever. Knowing how, it is a very powerfull program. Then, you might use ethereal, and set it up with a filter for smtp and domain. Depending on your system, it might by a big file. Kind of brute force method, I'm afraid. -- Cheers, Carlos Robinson