[Mailman-Users] Very busy list says send_digests failed, Too many links (fwd)

Mark Sapiro mark at msapiro.net
Fri Aug 28 18:08:46 CEST 2009


Ivan Fetch wrote:
>
>    I'm sorry to email you directly, but I believe the below email (sent 
>twice) is not making it to the mailman-users list. I just checked archives 
>and don't see the email there either. My re-post from yesterday, seems to 
>have been accepted by the python.org mail server:
>
>Aug 27 16:20:23 eros postfix/qmgr[13538]: [ID 197553 mail.info] 169B2C02B: 
>from=<ifetch at du.edu>, size=1764, nrcpt=1 (queue active)
>Aug 27 16:20:29 eros postfix/smtp[17110]: [ID 197553 mail.info] 169B2C02B: 
>to=<mailman-users at python.org>, relay=mail.python.org[82.94.164.166]:25, 
>delay=6.7, delays=0.01/0/5.9/0.73, dsn=2.0.0, status=sent (250 2.0.0 Ok: 
>queued as A1B94E309)
>Aug 27 16:20:29 eros postfix/qmgr[13538]: [ID 197553 mail.info] 169B2C02B: 
>removed


I suspect you are moderated as a recent subscriber and the moderator
just hasn't approved the latest posts. This happens.


>    Have you seen the below - on the mailman-users list, or in a queue 
>somewhere?


No, I haven't seen it, but I can't see the messages waiting moderation.
But, see below for my response.


>---------- Forwarded message ----------
>Date: Thu, 27 Aug 2009 16:19:58 -0600 (MDT)
>From: Ivan Fetch <ifetch at du.edu>
>To: mailman-users at python.org
>Subject: Very busy list says send_digests failed, Too many links
>
>Apologies - this is a re-post, but I didn't see it hit the list in about 19 
>hours.
>
>
>Date: Wed, 26 Aug 2009 21:14:00 -0600 (MDT)
>From: Ivan Fetch <ifetch at du.edu>
>To: mailman-users at python.org
>Subject: Very busy list says send_digests failed, Too many links
>
>Hello,
>
>
>    We had a list get around 30,000 messages today. Eventually, we noticed 
>emails to this list were really piling up in Mailman's "in" queue, and the 
>error log showed a lot of:
>
>Aug 26 13:44:22 2009 (9443) send_digests() failed: [Errno 31] Too many links: 
>'/mail/mailman/mailman/archives/private/list-name/attachments/20090826/9a8f3979'
>
>    I happened to get around this by setting digestable to know, but would like 
>to know how I should perhaps clean up after this. Most of the emails to this 
>list, unfortunately were the result of a run-away program sending emails. Will 
>the next digest cron job die on this many emails and attachments?


Yes, the next cron/senddigests will probably choke too unless you set
the list's digest_send_periodic to No.

The problem is in the list's lists/LISTNAME/digest.mbox file.
Apparently you have a very large number of messages with attachments
and with the same Message-ID:. Scrubber is calculating the directory
in which to store attachments for this message as
/mail/mailman/mailman/archives/private/list-name/attachments/20090826/9a8f3979
(the 9a8f3979 is a hash of the Message-ID:). This directory has so
many files in it that the OS will not allow any more to be created -
thus, the error 31.

You need to either edit the lists/LISTNAME/digest.mbox file and delete
all the 'bad' messages, or if you're not concerned about the digest,
just remove it. You also should remove the
archives/private/list-name/attachments/20090826/9a8f3979 directory.
That should be all the cleanup you need to do other that ensuring that
the run-away program is fixed.

-- 
Mark Sapiro <mark at msapiro.net>        The highway is for gamblers,
San Francisco Bay Area, California    better use your sense - B. Dylan



More information about the Mailman-Users mailing list