mailing lists and "unknown mail transport error"
in [Postfix]
|
Prev: dnswl doesn't work?
Next: why no configure script?
From: Dominik Storck on 9 Aug 2010 17:52 Hello Wietse, I have been tracking down to "Automatically tracing a Postfix daemon process" from the DEBUG_README. This is how I got the log snippet. Can't reproduce the error now after deleting unnecessary .forward files. Hope things will stay this way. Thank you Dominik Am 09.08.2010 22:17, schrieb Wietse Venema: > In case you wonder, I wrote Postfix. > > Perhaps you can follow instructions in > > http://www.postfix.org/.DEBUG_README.html#logging > > TURN OFF -v logging before you do this. > > Wietse
From: Wietse Venema on 9 Aug 2010 18:07 Dominik Storck: > Hello Wietse, > > I have been tracking down to "Automatically tracing a Postfix daemon > process" from the DEBUG_README. This is how I got the log snippet. > > Can't reproduce the error now after deleting unnecessary .forward files. Please do not waste your time with verbose logging. Please do not waste your time with debuggers, LONG BEFORE the queue manager defers mail with "unknown transport error", some other part of Postfix will log a WARNING:, FATAL: or PANIC: record with the reason why the program is failing. This is why I point you to DEBUG_README.html#logging and not to the other DEBUG_README sections. Wietse
From: Dominik Storck on 10 Aug 2010 01:33 Hello Wietse, I have been looking for these error messages over and over before I started digging deeper. There are none of the error,fatal,warning or panic messages before. The first error log entry to occur is the unknown mail transport error, really. Dominik Am 10.08.2010 00:07, schrieb Wietse Venema: > Dominik Storck: >> Hello Wietse, >> >> I have been tracking down to "Automatically tracing a Postfix daemon >> process" from the DEBUG_README. This is how I got the log snippet. >> >> Can't reproduce the error now after deleting unnecessary .forward files. > > Please do not waste your time with verbose logging. > > Please do not waste your time with debuggers, > > LONG BEFORE the queue manager defers mail with "unknown transport > error", some other part of Postfix will log a WARNING:, FATAL: or > PANIC: record with the reason why the program is failing. > > This is why I point you to DEBUG_README.html#logging and not > to the other DEBUG_README sections. > > Wietse
From: Wietse Venema on 10 Aug 2010 07:35 Dominik Storck: > Hello Wietse, > > I have been looking for these error messages over and over before I > started digging deeper. There are none of the error,fatal,warning or > panic messages before. The first error log entry to occur is the unknown > mail transport error, really. Looking at the original problem report: Aug 2 14:52:06 postfix postfix/local[29097]: 720441D34008: to=<wie(a)postfix.meinedomain.com>, orig_to=<all(a)meinedomain.com>, relay=local, delay=77, delays=0.23/0/0/77, dsn=2.0.0, status=sent (delivered to command: /usr/bin/vacation -j wie) Aug 2 14:52:06 postfix postfix/local[29097]: 720441D34008: to=<wol(a)postfix.meinedomain.com>, orig_to=<all(a)meinedomain.com>, relay=local, delay=77, delays=0.23/0/0/77, dsn=2.0.0, status=sent (delivered to maildir) Aug 2 14:52:06 postfix postfix/local[29097]: 720441D34008: to=<wre(a)postfix.meinedomain.com>, orig_to=<all(a)meinedomain.com>, relay=local, delay=77, delays=0.23/0/0/77, dsn=2.0.0, status=sent (delivered to maildir) Aug 2 14:52:08 postfix postfix/error[32270]: 720441D34008: to=<all(a)meinedomain.com>, relay=none, delay=79, delays=0.23/79/0/0.67, dsn=4.3.0, status=deferred (unknown mail transport error) When you have a many-recipient "all" alias, you need to set up an "owner-all" alias in the alias database (with the right-hand side being the adminstrator of the "all" list). Besides changing the way errors are reported, this also triggers a different delivery strategy where one slow recipient won't cause all deliveries to be deferred. Wietse
From: "Dominik Storck" on 10 Aug 2010 08:52
Hello Wietse, have not been aware of the difference. Added the owner-alias and will keep an eye on the list. Thanks a lot Dominik > Dominik Storck: >> Hello Wietse, >> >> I have been looking for these error messages over and over before I >> started digging deeper. There are none of the error,fatal,warning or >> panic messages before. The first error log entry to occur is the unknown >> mail transport error, really. > > Looking at the original problem report: > > Aug 2 14:52:06 postfix postfix/local[29097]: 720441D34008: > to=<wie(a)postfix.meinedomain.com>, orig_to=<all(a)meinedomain.com>, > relay=local, delay=77, delays=0.23/0/0/77, dsn=2.0.0, status=sent > (delivered to command: /usr/bin/vacation -j wie) > Aug 2 14:52:06 postfix postfix/local[29097]: 720441D34008: > to=<wol(a)postfix.meinedomain.com>, orig_to=<all(a)meinedomain.com>, > relay=local, delay=77, delays=0.23/0/0/77, dsn=2.0.0, status=sent > (delivered to maildir) > Aug 2 14:52:06 postfix postfix/local[29097]: 720441D34008: > to=<wre(a)postfix.meinedomain.com>, orig_to=<all(a)meinedomain.com>, > relay=local, delay=77, delays=0.23/0/0/77, dsn=2.0.0, status=sent > (delivered to maildir) > > Aug 2 14:52:08 postfix postfix/error[32270]: 720441D34008: > to=<all(a)meinedomain.com>, relay=none, delay=79, delays=0.23/79/0/0.67, > dsn=4.3.0, status=deferred (unknown mail transport error) > > When you have a many-recipient "all" alias, you need to set up an > "owner-all" alias in the alias database (with the right-hand side > being the adminstrator of the "all" list). Besides changing the > way errors are reported, this also triggers a different delivery > strategy where one slow recipient won't cause all deliveries to be > deferred. > > Wietse > |