[Mailman-Users] undelivered mail, but only to some

[D. Alan Stewart]

Responding to Mark Sapiro:

> If the sendmail log says the post is sent to the user in question,
> Mailman delivered it to sendmail for the user, and it's not Mailman's
> problem.

I agree, it's not clear whether this is a sendmail or a Mailman  
problem. However, I have never, AFAIK, failed to receive an email  
from my provider's domain, unless that email originated from this  
specific Mailman list. That seems suspicious, doesn't it? Of course,  
I may be missing other emails that I am not aware of. Not getting  
something you did not know to expect is not an easy problem to trap.

Since I now know that I am not receiving all emails from this list,  
possibly the problem is common but not often detected because the  
severity varies, list by list.

> Mailman's 'smtp' log. This will tell you how many recipients the post
> was sent to.

My provider has checked Mailman's smtp log. It shows emails were sent  
to all subscribers.

> Also check if the list is personalized. You may not even have
> personalization options on your Non-digest options page which means
> that OWNERS_CAN_ENABLE_PERSONALIZATION is not set to Yes in mm_cfg.py
> and delivery is almost certainly not personailzed.

I'm not certain, but I think this answers your question: the list  
does have digest options available to the users.

> Check the settings of VERP_DELIVERY_INTERVAL and SMTP_MAX_RCPTS if any
> in mm_cfg.py

I looked, there are no settings for these in Mailman/mm_cfg.py. There  
don't appear to be any site-specific settings here.

> You should also see this in the sendmail log, i.e. that the mail was
> delivered to sendmail in one transaction.

The provider has given me sendmail logs that I have analyzed. I was  
able to identify entries in the sendmail log that correspond with  
mailing list transmissions that were not received by the subscriber  
(me).

> Sendmail claims it sent the mail to the users who didn't get it. I
> recognize the symptoms say that Mailman is somehow involved, but I
> don't see how.

I recognize the difficulty of this particular problem.

Thanks,
Alan