Christian Stalberg via Mailman-users writes:
I went into nginx.conf and raised the client_max_body_size from the default 1M to 0, then restarted nginx. The '413 Request Entity Too Large' error has gone away, but postings to the list are not distributed. The mail queue is empty. Now what?
How do you know the posts aren't distributed? If it's posters using Google mail complaining about not seeing their own posts, tell them to look in their Sent Mail folders.
If your HTTPd was configured with a 1MB size limit, how about your MTA(s) and spam/malware filters? Has the mail spool's storage filled up?
What is the mail that is not being delivered? If you configure Mailman to strip HTML, it's possible that the messages were empty after stripping an HTML part, in which case Mailman ignores them. In that case, try switching from strip HTML to convert HTML to plain text, or allow HTML.
Otherwise, what is your MTA software? What mail filters (spam, size limits) are configured into it? What do you know about the affected posters, in particular, what are their systems' DMARC policies?
Is the mail coming from inside your organization and passing through only your MTAs, or are there external MTAs involved? Is Mailman's MTA a boundary MTA (the MTA that talks to systems on the Internet outside your organization)? If not, what is in the logs for the boundary system?
What do Mailman's MTA's logs say? Did the posts get to the MTA from wherever? Does the MTA say Mailman accepted them? Mailman can be a bit spotty about logging success, but errors are not supposed to pass silently -- if there was a failure Mailman knows about it should be logged in the Mailman logs. Then, what does the MTA log say about outgoing mail?
Steve