If you look in your Deferred queue, you may notice a number of emails that did not get sent due to connection timeout errors.
Postfix themselves recommend setting short(er) connection timeouts to avoid wasting time/resources trying to send email to overloaded servers.
But, for whatever reasons, I was noticing that my own Deferred queues have been over the past year increasing in average size, so I did an experiment. I increased two timeouts by running (as Zimbra):
Some 3/4 of the previously deferred emails were delivered. The remainder were almost all "over quota" messages. I'm keeping the above settings.
Since there are no LDAP attributes for these Postfix parameters (see https://wiki.zimbra.com/wiki/Postconf_keys for the full list), you may need to re-apply these changes after a patch/upgrade.
Further, the above attribute values may not be optimal for your environment. Best to document what they are now, increase gradually, and see how that impacts your Deferred queue.
Hope that helps,
Mark
Postfix themselves recommend setting short(er) connection timeouts to avoid wasting time/resources trying to send email to overloaded servers.
But, for whatever reasons, I was noticing that my own Deferred queues have been over the past year increasing in average size, so I did an experiment. I increased two timeouts by running (as Zimbra):
Code:
postconf -e 'smtp_connect_timeout = 300s'postconf -e 'smtp_helo_timeout = 120s'postfix reloadpostfix flushSince there are no LDAP attributes for these Postfix parameters (see https://wiki.zimbra.com/wiki/Postconf_keys for the full list), you may need to re-apply these changes after a patch/upgrade.
Further, the above attribute values may not be optimal for your environment. Best to document what they are now, increase gradually, and see how that impacts your Deferred queue.
Hope that helps,
Mark
Statistics: Posted by L. Mark Stone — Mon Feb 10, 2025 5:33 pm