Re: [Mailman-Users] tuning bounce behavior with VERP

2013-09-12 Thread Mark Sapiro 
On 09/12/2013 08:34 AM, Perry E. Metzger wrote:
> 
> So I'm having some trouble figuring out whether things are working
> correctly. In particular, I seem to have users who are undeliverable
> for weeks and yet don't seem to be being unsubed even though the
> settings should be doing that.


If their delivery is disabled by bounce and not for some other reason,
and they are not being removed, the most likely explanation is
cron/disabled is not being run.

If they are not getting delivery disabled, check Mailman's 'bounce' log.
Also check the Postfix logs to see if bounces are returned and where
they are delivered.


> Is there a good way to examine the database of bounce information
> being stored so that I can tell what Mailman things the state for
> such users is? 


See, 

-- 
Mark Sapiro The highway is for gamblers,
San Francisco Bay Area, Californiabetter use your sense - B. Dylan
--
Mailman-Users mailing list Mailman-Users@python.org
https://mail.python.org/mailman/listinfo/mailman-users
Mailman FAQ: http://wiki.list.org/x/AgA3
Security Policy: http://wiki.list.org/x/QIA9
Searchable Archives: http://www.mail-archive.com/mailman-users%40python.org/
Unsubscribe: 
https://mail.python.org/mailman/options/mailman-users/archive%40jab.org

Re: [Mailman-Users] tuning bounce behavior with VERP

2013-09-12 Thread Perry E. Metzger 
On Mon, 02 Sep 2013 11:50:48 -0700 Mark Sapiro 
wrote:
> On 09/02/2013 11:35 AM, Barry S. Finkel wrote:
> 
> > A fatal NDR (once per day) is
> > treated as a bounce, and increases the bounce score by 1.0 . A
> > soft bounce is counted as 0.5 .
> 
> 
> That's how it's documented, but it doesn't actually work that way. A
> bounce is either fatal (scored as 1.0) or it is ignored. Nothing is
> scored as 0.5.
> 

So I'm having some trouble figuring out whether things are working
correctly. In particular, I seem to have users who are undeliverable
for weeks and yet don't seem to be being unsubed even though the
settings should be doing that.

Is there a good way to examine the database of bounce information
being stored so that I can tell what Mailman things the state for
such users is? 

Perry
-- 
Perry E. Metzgerpe...@piermont.com
--
Mailman-Users mailing list Mailman-Users@python.org
https://mail.python.org/mailman/listinfo/mailman-users
Mailman FAQ: http://wiki.list.org/x/AgA3
Security Policy: http://wiki.list.org/x/QIA9
Searchable Archives: http://www.mail-archive.com/mailman-users%40python.org/
Unsubscribe: 
https://mail.python.org/mailman/options/mailman-users/archive%40jab.org

Re: [Mailman-Users] tuning bounce behavior with VERP

2013-09-02 Thread Mark Sapiro 
On 09/02/2013 02:26 PM, Perry E. Metzger wrote:
> 
> I'm more interested in questions like how I tell Mailman that I'm
> using VERP at all and such. Anything that is on the bounce
> configuration page I can just do through the web interface. :)


Your OP implied that you were doing the actual VERP in Postfix. If that
is the case, it should 'just work' without any Mailman changes at all as
long as your VERPed envelope sender for the LISTNAME list message to
u...@example.com looks like

  LISTNAME-bounces+user=example.com@list.domain

I.e., Postfix VERP delimiters are '+='. This is parsed by

VERP_REGEXP =
r'^(?P[^+]+?)\+(?P[^=]+)=(?P[^@]+)@.*$'


If your VERPed envelope sender is different, you will need to set
VERP_REGEXP in mm_cfg.py to parse it correctly.

Also see the FAQ at .

-- 
Mark Sapiro The highway is for gamblers,
San Francisco Bay Area, Californiabetter use your sense - B. Dylan
--
Mailman-Users mailing list Mailman-Users@python.org
http://mail.python.org/mailman/listinfo/mailman-users
Mailman FAQ: http://wiki.list.org/x/AgA3
Security Policy: http://wiki.list.org/x/QIA9
Searchable Archives: http://www.mail-archive.com/mailman-users%40python.org/
Unsubscribe: 
http://mail.python.org/mailman/options/mailman-users/archive%40jab.org

Re: [Mailman-Users] tuning bounce behavior with VERP

2013-09-02 Thread Perry E. Metzger 
On Mon, 02 Sep 2013 11:50:48 -0700 Mark Sapiro 
wrote:
> On 09/02/2013 11:35 AM, Barry S. Finkel wrote:
> 
> > A fatal NDR (once per day) is
> > treated as a bounce, and increases the bounce score by 1.0 . A
> > soft bounce is counted as 0.5 .
> 
> That's how it's documented, but it doesn't actually work that way. A
> bounce is either fatal (scored as 1.0) or it is ignored. Nothing is
> scored as 0.5.

I'm more interested in questions like how I tell Mailman that I'm
using VERP at all and such. Anything that is on the bounce
configuration page I can just do through the web interface. :)

Perry
-- 
Perry E. Metzgerpe...@piermont.com
--
Mailman-Users mailing list Mailman-Users@python.org
http://mail.python.org/mailman/listinfo/mailman-users
Mailman FAQ: http://wiki.list.org/x/AgA3
Security Policy: http://wiki.list.org/x/QIA9
Searchable Archives: http://www.mail-archive.com/mailman-users%40python.org/
Unsubscribe: 
http://mail.python.org/mailman/options/mailman-users/archive%40jab.org

Re: [Mailman-Users] tuning bounce behavior with VERP

2013-09-02 Thread Mark Sapiro 
On 09/02/2013 11:35 AM, Barry S. Finkel wrote:

> A fatal NDR (once per day) is
> treated as a bounce, and increases the bounce score by 1.0 . A soft
> bounce is counted as 0.5 .


That's how it's documented, but it doesn't actually work that way. A
bounce is either fatal (scored as 1.0) or it is ignored. Nothing is
scored as 0.5.

-- 
Mark Sapiro The highway is for gamblers,
San Francisco Bay Area, Californiabetter use your sense - B. Dylan
--
Mailman-Users mailing list Mailman-Users@python.org
http://mail.python.org/mailman/listinfo/mailman-users
Mailman FAQ: http://wiki.list.org/x/AgA3
Security Policy: http://wiki.list.org/x/QIA9
Searchable Archives: http://www.mail-archive.com/mailman-users%40python.org/
Unsubscribe: 
http://mail.python.org/mailman/options/mailman-users/archive%40jab.org

Re: [Mailman-Users] tuning bounce behavior with VERP

2013-09-02 Thread Barry S. Finkel 

On 9/2/2013 12:52 PM, Perry E. Metzger wrote:

I have a Postfix setup with VERP turned on for some lists that I run.
(I'm cool with the load using VERP with every message implies.)

I'd like to be able to tune bounce behavior, but the documentation for
this seems a bit spotty. What can I adjust to make things behave the
way I might like?

Perry
-- Perry E. Metzger pe...@piermont.com


If I remember all of the parameters correctly, you can set

a) the bounce score when the bounce is determined "fatal"
   and the NDR message is sent to the list owner(s).
   The default is 5.0 .

I do not remember the other settings.  But the URL

 http://www.gnu.org/software/mailman/mailman-admin/node25.html

lists them.  A fatal NDR (once per day) is
treated as a bounce, and increases the bounce score by 1.0 . A soft
bounce is counted as 0.5 .  When I was running a Mailman server I ran
a daily report that gave me for each list all of the subscribers with
a bounce score greater than 0.0; this was via a script written by Marc
Sapiro.

The list owner(s) does/do not see any NDR until the bounce score reaches
the maximum.

--Barry Finkel
--
Mailman-Users mailing list Mailman-Users@python.org
http://mail.python.org/mailman/listinfo/mailman-users
Mailman FAQ: http://wiki.list.org/x/AgA3
Security Policy: http://wiki.list.org/x/QIA9
Searchable Archives: http://www.mail-archive.com/mailman-users%40python.org/
Unsubscribe: 
http://mail.python.org/mailman/options/mailman-users/archive%40jab.org

[Mailman-Users] tuning bounce behavior with VERP

2013-09-02 Thread Perry E. Metzger 
I have a Postfix setup with VERP turned on for some lists that I run.
(I'm cool with the load using VERP with every message implies.)

I'd like to be able to tune bounce behavior, but the documentation for
this seems a bit spotty. What can I adjust to make things behave the
way I might like?

Perry
-- 
Perry E. Metzgerpe...@piermont.com
--
Mailman-Users mailing list Mailman-Users@python.org
http://mail.python.org/mailman/listinfo/mailman-users
Mailman FAQ: http://wiki.list.org/x/AgA3
Security Policy: http://wiki.list.org/x/QIA9
Searchable Archives: http://www.mail-archive.com/mailman-users%40python.org/
Unsubscribe: 
http://mail.python.org/mailman/options/mailman-users/archive%40jab.org