Message-sending delays to federated relays because deferred mail queue is full

maren-bunk · May 4, 2026, 5:58pm

hello! i’m one of the adminstrators of the https://sweetfern.net relay.

we’re having an issue where messages to other relays take multiple hours to deliver. this issue was discovered this morning, and i think that it’s because the postfix mail queue is full of messages waiting to get sent to misconfigured/blocked relays that they will never successfully deliver to.

from the postfix@-.service logs:

May 04 17:42:09 sweetfern postfix/qmgr[1168]: warning: mail for inet:[127.0.0.1]:10083 is using up 12203 of 12203 active queue entries
May 04 17:42:09 sweetfern postfix/qmgr[1168]: warning: you may need to increase the main.cf lmtp-filtermail_destination_concurrency_limit from 20
May 04 17:42:09 sweetfern postfix/qmgr[1168]: warning: please avoid flushing the whole queue when you have
May 04 17:42:09 sweetfern postfix/qmgr[1168]: warning: lots of deferred mail, that is bad for performance
May 04 17:42:09 sweetfern postfix/qmgr[1168]: warning: to turn off these warnings specify: qmgr_clog_warn_time = 0

running postqueue -p shows that the queue is in fact extremely full.

i’d like to mitigate this issue manually if possible, because sweetfern is currently not able to reasonably send messages to other relays as a result. is there a way to safely expedite the queue flushing process, without losing those messages stuck in queue that are trying to get sent to correctly-configured servers?

i’m assuming that purging the queue will cause lost messages, because some of these messages will get delivered once it’s their turn in the queue. but the logs instruct not to flush the whole queue. i’m going to try flushing the queue despite the logs instructions to avoid flushing the whole queue, but any other tips or tricks would be much appreciated <3

link2xt · May 5, 2026, 12:32am

This is likely caused by the recent change on the main branch:

This is not in 1.10.0 release, only on main. If you deployed from main, you can disable it by opening /etc/postfix/main.cf and commenting out the last line starting with default_transport:

github.com/chatmail/relay

cmdeploy/src/cmdeploy/postfix/main.cf.j2

44fe2dc08


      
          
          # Do not lookup SMTP client hostnames to reduce delays
          # and avoid unnecessary DNS requests.
          smtpd_peername_lookup = no
          
          # Use filtermail-transport to relay messages.
          # We can't force postfix to split messages per destination,
          # when specifying a custom next-hop,
          # so instead this is handled in filtermail.
          # We use LMTP instead SMTP so we can communicate per-recipient errors back to postfix.
          default_transport = lmtp-filtermail:inet:[127.0.0.1]:{{ config.filtermail_lmtp_port_transport }}

maren-bunk · May 5, 2026, 1:42am

this fixed it immediately, thank you SO much.