Hi Trying to avoid my IP being grey/blacklisted (mainly with microsoft domains: hotmail/outlook/live.com), I've put the following option in my postfix main.cf default_destination_rate_delay = 10s This is working great, and every other mail sent to the same domain gets delayed for 10s by postfix. This is what I need for not being rate limited by microsoft But mailqueue started to get increased and it's because with this option active, only one of the eight amavis daemons I have up and ready are used to process mail. Only one avavis process cannot cope with the normal traffic in my server. As soon as I comment out the option and restart postfix, mail system starts to use 8 amavis process again. I've tried with "1s" instead of 10s, with no difference. Only one simultaneous amavis process is used when rate delay is active What am I missing that is making postfix ignore amavis settings when I activate rate delay? BTW, this is my amavis configuration for increasing process: grep servers /etc/amavis/conf.d/51-user-customised $max_servers = 8; grep amavis /etc/postfix/master.cf amavis unix - - - - 8 smtp Thank you!
Try to set the delay option for the relevant transport (smtp) in master.cf only instead of setting it globally in main.cf
I've put it this way: smtp inet n - - - - smtpd -o default_destination_rate_delay=10s But it's not working at all. I can send as many mails to a domain as I want without any throttle.
Hmm, ok. It was worth a try. You can either try to set it for smtps and submission as well. if this does not work, you can try to set it in main.cf globally again and set it to 0 for the amavis and 10025 and 10027 port transports.
Tried adding it to "smtps" and "submission", and it made no difference. Putting options in the amavis transports makes mail system fail: May 18 02:17:05 HOST amavis[18656]: (18656-11) (!)rw_loop: leaving rw loop, no progress, last event (select) 35.030 s ago May 18 02:17:05 HOST amavis[18656]: (18656-11) (!)mfJ2MztC4YMc FWD from <NAME@OR_DOMAIN.com> -> <NAME1@DEST_DOMAIN.com>,<NAME2a@DEST_DOMAIN.com>,<NAME3@DEST_DOMAIN.com>, 451 4.5.0 From MTA() during fwd-connect (No greeting, dt: 35.030 s): id=18656-11 May 18 02:17:05 HOST amavis[18656]: (18656-11) Blocked MTA-BLOCKED {TempFailedOutbound}, ORIGINATING LOCAL [213.97.253.195]:47014 [213.97.253.195] <NAME@OR_DOMAIN> -> <NAME1@DEST_DOMAIN.com>,<NAME2a@DEST_DOMAIN.com>,<NAME3@DEST_DOMAIN.com>, Queue-ID: 66DF0B478DE, Message-ID: <6263951.DYVVNObH9c@elle>, mail_id: mfJ2MztC4YMc, Hits: -2.9, size: 1658, 92283 ms Deleting it solves the problem. Any other idea?
It seems the last error was a transient error. The configuration works, but it seems it makes postfix use only one amavis process again. Any other idea?
Not really, never tried to configure it like this. Maybe someone else has an idea how to do that or you might want to ask @florian030 if he knows how to set it up so that all amavis processes are used. What I did in the past to get email sending quotas was to use policyd (cluebringer), but that's a few years ago and I don't have a ready made guide for you on that topic.
