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?
-----Original Message----- From: Stephen J. Turnbull <stephenjturnbull@gmail.com> Sent: Tuesday, October 19, 2021 7:20 AM To: Christian Stalberg <csa@web-analysts.net> Cc: mailman-users@mailman3.org Subject: [MM3-users] Request Entity Too Large
Christian Stalberg via Mailman-users writes:
Now mail sent to mm3 for distribution is not going out, for two > different lists.
Check whether stuff is stuck in your MTA queue, and your MTA logs. Also check Mailman's queues (usually in /var/lib/mailman/qfiles, I think).
As for why it's two different lists, is it possible that somebody cross-posted to those lists, or one is subscribed to the other?
The mailman.log shows this error:
Oct 17 21:50:00 2021 (491) ACCEPT: <030601d7c3a0$ee342f20$ca9c8d60$@somedomain.com> Oct 17 21:50:04 2021 (488) HyperKitty failure on > https://somedomain.net/archives/api/mailman/archive: <html>
I don't think this should affect mail going out in itself, because this appears to be delivery to HyperKitty, not to any remote host. Why nginx is involved is a mystery to me, though.
<head><title>413 Request Entity Too Large</title></head> > <body bgcolor="white"> > <center><h1>413 Request Entity Too Large</h1></center> <hr><center>nginx/1.14.2</center> > </body> > </html> > (413) > Oct 17 21:50:04 2021 (488) Exception in the HyperKitty archiver: <html> > <head><title>413 Request Entity Too Large</title></head> > <body bgcolor="white"> > <center><h1>413 Request Entity Too Large</h1></center> <hr><center>nginx/1.14.2</center> > </body> > </html>
It appears that nginx is complaining that a request from HyperKitty is too large. I don't know much about nginx, but usually a request is sent FROM a client TO an HTTPd, and it's saying it's too big. I guess if nginx is the front end for HyperKitty, it could be that some client is requesting a huge email from HyperKitty.
I guess you could have nginx between Mailman core and HyperKitty, in which case this is apparently an nginx configuration issue: it has an upper limit on the size of messages it's willing to accept from core and pass on to HyperKitty.
If your Mailman is getting messages large enough that an HTTPd balks, I wouldn't be surprised if your MTA is on strike, or is getting "you gotta be kidding, I'm too busy, try tomorrow" from remote recipients.
It could be a Mailman problem, but there's nothing in your report to suggest that it is.