I want to thank you for pointing us in the right direction, and also for taking the time to answer my plea for help.
To answer your questions: 1. No, I don't have proof positive that a buffer overrun was the cause of my troubles, but it seems to fit the situation too well to not be. What I didn't realize when I originally posted my call for help is that it also had a fairly large image attached as well. Between the formatting, the file, and the CC's I'm relatively sure that it was sufficient to force the overrun. I will ask the admin for the logs, and forward them on to you privately, if that's OK. If we've uncovered a bug, I definitely want to help improve things! 2. I'm not aware of any custom config-work in this implementation. It's a standard mailman running on BSD, with just a few lists - the other lists were untouched by this, fortunately! It may have been the admin's attempt to 'dumb it down' for me that makes this confusing after the fact. Your suggestion to review the postfix settings got me looking in the right place. Long and short of it: I was able to use postfixadmin to delete and the re-create the aliases in postfix, at which point, everything started running as normal. Now for the suggestion/feature request(s): 1 - would it be possible to add a "show all list members" view, and sort them by their status in each column? Actually, just being able to sort any given letter by their various status settings would be tremendous - all no mails, all no duplicates, etc in a row. 2 - What would be an incredibly powerful tool to the non-technical list admin would be the ability to download a text file of listmembers and their status settings. just a simple txt file, showing name, email address, and their basic digest/no mail/ moderated/plain text settings would be a wonderful thing. I'd be thrilled if just the name and email address would be available this way! Thanks, Perry ----- Perry M Lynch, CISSP CISA JNCIA-FWV, RSA CSE (SecurID), CCNA "why yes, I am certifiable..." ----- ----- Original Message ---- From: Mark Sapiro <[EMAIL PROTECTED]> To: Perry M Lynch <[EMAIL PROTECTED]>; mailman-users@python.org Sent: Saturday, March 1, 2008 11:18:45 AM Subject: Re: [Mailman-Users] In need of help. Trying to recover from a seriousbuffer overflow Perry M Lynch wrote: > >On Wednesday, a member submitted a large email with poor formatting to the >list. And it had enough CCs attached to it that it caused a buffer overflow, >which resulted in the message being self-approved for the list. Do you have evidence that this is what happened, or is this just conjecture? >Suddenly the server began repeating itself as it tried to process this huge >message over and over again. It became a vicious cycle - as the one message >finally repeated out, other messages that were caught in the queue became >ammunition for the server to repeat. Prior to the system crash, more than 6100 >duplicate messages were sent to 450 members, at an estimated peak rate of >about 20,000 outbound messages per half hour. > >The user db became corrupt, and my server admin and I have worked to get >things operational again. At this point, the list is supposedly behaving >itself, yet no messages are actually being sent from the server. I don't have >access to the raw logs, Which is unfortunate, because that's what we need to see to determine what happened. I would like to see Mailman's error, post, qrunner, smtp, smtp-failure and vette logs from the time of the original failure to try to see what happened and if there is a Mailman bug of some kind that allowed it. Can you get these logs from the server admin? >but here is the snippet of discussion about the restoration of the user >database. > >Jason: here's what I see on the server. > > # pwd > /var/spool/mailman/lists/cyberscots > # file * > config.pck: data > config.pck.last: data > digest.mbox: UTF-8 Unicode mail text > en: setgid directory > pending.pck: ASCII text > request.pck: data > >and the restore at home... > > /backups/spool/mailman/lists/cyberscots > # file * > config.pck: smtp mail text > config.pck.last: smtp mail text > digest.mbox: ISO-8859 mail text > en: directory > pending.pck: smtp mail text > request.pck: data > >the *.pck files from the restore all look like regular mbox files. Which is totally wrong, because they should be Python pickles. >but the ones on the server are some sort of binary config file. Which is what they should be. >I checked against a different list, same thing. >on the server... > /var/spool/mailman/lists/lynches > # file * > config.pck: data > config.pck.last: data > >we got a user db restored, What user db? In standard Mailman, the user db, list config, etc. is all part of the list object saved in the config.pck. Do you have a custom MemberAdaptor that uses a different, separate user db? or are you just referring to the config.pck? >at which point the gui became functional again, and we cleared the outbound >queue, Mailman's out queue or the Postfix queue? >which had 1400+ messages that were blocked by a provider who'd had enough. And >now, although I can send emails to the list, they do not seem to make it all >the way through the process: Is Mailman running? See <http://www.python.org/cgi-bin/faqw-mm.py?req=show&file=faq04.078.htp>. >Feb 29 22:15:10 colo2 amavis[27834]: (27834-18) ESMTP::10024 >/var/amavisd/tmp/amavis-20080229T194128-27834: <[EMAIL PROTECTED]> -> <[EMAIL >PROTECTED]> Received: SIZE=3425 from mx1.dixongroup.net ([127.0.0.1]) by >localhost (mx1.dixongroup.net [127.0.0.1]) (amavisd-new, port 10024) with >ESMTP id 27834-18 for <[EMAIL PROTECTED]>; Fri, 29 Feb 2008 22:15:10 -0500 >(EST) >Feb 29 22:15:10 colo2 amavis[27834]: (27834-18) Checking: f8-BzF9WBjIQ <[EMAIL >PROTECTED]> -> <[EMAIL PROTECTED]> >Feb 29 22:15:11 colo2 amavis[27834]: (27834-18) FWD via SMTP: <[EMAIL >PROTECTED]> -> <[EMAIL PROTECTED]>, 250 2.6.0 Ok, id=27834-18, from >MTA([127.0.0.1]:10025): 250 2.0.0 Ok: queued as 2177591EC2 >Feb 29 22:15:11 colo2 amavis[27834]: (27834-18) Passed CLEAN, [71.178.250.158] ><[EMAIL PROTECTED]> -> <[EMAIL PROTECTED]>, Message-ID: <[EMAIL PROTECTED]>, >mail_id: f8-BzF9WBjIQ, Hits: -, 671 ms >Feb 29 22:15:11 colo2 postfix/smtp[11051]: 7D8FD91EC1: to=<[EMAIL PROTECTED]>, >orig_to=<[EMAIL PROTECTED]>, relay=127.0.0.1[127.0.0.1]:10024, delay=0.97, >delays=0.28/0/0.01/0.68, dsn=2.6.0, status=sent (250 2.6.0 Ok, id=27834-18, >from MTA([127.0.0.1]:10025): 250 2.0.0 Ok: queued as 2177591EC2) >Feb 29 22:15:11 colo2 postfix/local[406]: 2177591EC2: to=<[EMAIL PROTECTED]>, >relay=local, delay=0.71, delays=0.21/0.1/0/0.4, dsn=2.0.0, status=sent >(delivered to command: /usr/local/lib/mailman/mail/mailman post cyberscots) OK. The post was delivered to the Mailman wrapper. <snip> > >In short, it looks as if messages are being received, and they are being >delivered to the outbound queue, but nothing is happening beyond that point. Actually, they are being delivered to Mailman's mail wrapper which terminates normally. This means they have been queued in Mailman's incoming queue. See the FAQ linked above for the next steps (you can skip steps 2 and 3 because this is Postfix, and we know delivery to Mailman is working). -- Mark Sapiro <[EMAIL PROTECTED]> The highway is for gamblers, San Francisco Bay Area, California better use your sense - B. Dylan ____________________________________________________________________________________ Looking for last minute shopping deals? Find them fast with Yahoo! Search. http://tools.search.yahoo.com/newsearch/category.php?category=shopping ------------------------------------------------------ Mailman-Users mailing list Mailman-Users@python.org http://mail.python.org/mailman/listinfo/mailman-users Mailman FAQ: http://www.python.org/cgi-bin/faqw-mm.py Searchable Archives: http://www.mail-archive.com/mailman-users%40python.org/ Unsubscribe: http://mail.python.org/mailman/options/mailman-users/archive%40jab.org Security Policy: http://www.python.org/cgi-bin/faqw-mm.py?req=show&file=faq01.027.htp
