FAQ
Hallo!

This may not specifically be a mailman problem, but I'm noticing
strange behaviour in mailmain, as well as in general on my mail server.

Background: My CentOS 4 system ran an automatic 'yum' update two days
ago, and it updated many system files. I'm not an expert, but it looks
like the 'kernel' and many 'lib' files were updated. The system did not
'reboot', but during this process everal daemons were automatically
restarted, and my 'Mail Avenger' daemon just failed (which caused
temporary delay of incoming mail). I had to restart 'asmtpd' and for good
measure, restarted postfix.

Mailman seemed to 'lock up', and queued all outgoing messages until I did
a restart mid-yesterday. I got another complaint today, did another
restart, and once again the list messages are going out. But I have no
guarantee that it won't keep repeating this behaviour. Has anyone heard or
seen any explanation for behaviour like this?

It may be a postfix issue, even though postfix was *not* among the updated
programs. I'm noticing that I'm seeing multiple copies of *some* messages
from this list. They have the Received Lines showing they were sent
multiple times by the list serve, almost as if they had 4xx errors the
first time? Very strange.....

I don't really want to reboot the whole system, especially where the
stability of it is in doubt. Any advice welcome (anyone know a good,
active CentOS mailing list that might have an angle on this?).

Thanks!

- Charles

Search Discussions

  • Mark Sapiro at Apr 24, 2009 at 5:13 pm

    Charles Gregory wrote:
    Mailman seemed to 'lock up', and queued all outgoing messages until I did
    a restart mid-yesterday. I got another complaint today, did another
    restart, and once again the list messages are going out. But I have no
    guarantee that it won't keep repeating this behaviour. Has anyone heard or
    seen any explanation for behaviour like this?

    It may be a postfix issue, even though postfix was *not* among the updated
    programs. I'm noticing that I'm seeing multiple copies of *some* messages
    from this list. They have the Received Lines showing they were sent
    multiple times by the list serve, almost as if they had 4xx errors the
    first time? Very strange.....

    I don't really want to reboot the whole system, especially where the
    stability of it is in doubt. Any advice welcome (anyone know a good,
    active CentOS mailing list that might have an angle on this?).

    First of all, my .02 on rebooting - If you do a yum upgrade that
    updates "system files", you should reboot. If you don't, and there is
    some issue due to the upgrade, the next time you have an unplanned
    reboot due to a power failure or some other external event, you will
    have problems ultimately caused by this upgrade, and you may not even
    remember this upgrade as a potential cause.

    Regarding the Mailman issue per se, was Python involved in the upgrade?

    See various Mailman logs such as error and smtp-failure for problems.

    See the FAQ at <http://wiki.list.org/x/A4E9>.

    --
    Mark Sapiro <mark at msapiro.net> The highway is for gamblers,
    San Francisco Bay Area, California better use your sense - B. Dylan
  • Charles Gregory at Apr 24, 2009 at 6:24 pm

    On Fri, 24 Apr 2009, Mark Sapiro wrote:
    Mailman seemed to 'lock up', and queued all outgoing messages until I did
    a restart....
    First of all, my .02 on rebooting - If you do a yum upgrade that
    updates "system files", you should reboot. If you don't, and there is
    some issue due to the upgrade, the next time you have an unplanned
    reboot due to a power failure or some other external event, you will
    have problems ultimately caused by this upgrade, and you may not even
    remember this upgrade as a potential cause.
    My problem is that my system is in a hard-to-access co-lo, which means I
    don't want to risk a reboot that causes problems that can only be
    addressed from a console. So I'm hoping I can avoid the reboot. Otherwise
    I'll have to schedule it for when the box is accessible.....
    Regarding the Mailman issue per se, was Python involved in the upgrade?
    Don't think so.
    See various Mailman logs such as error and smtp-failure for problems.
    Interestingly enough, when I shutdown mailman it reports an SMTP failure
    for a message it appears to have been 'attempting' to send, then
    shuts down. If I have a 'bad' message in the queue, how would I check
    this, and erase/drop it?
    See the FAQ at <http://wiki.list.org/x/A4E9>.
    Took a run through. Most of the suggestions relate to 'it never worked'
    conditions, as opposed to 'it was working and stopped', or 'it only works
    when I restart it'.

    - C
  • Mark Sapiro at Apr 24, 2009 at 7:31 pm

    Charles Gregory wrote:
    On Fri, 24 Apr 2009, Mark Sapiro wrote:

    See various Mailman logs such as error and smtp-failure for problems.
    Interestingly enough, when I shutdown mailman it reports an SMTP failure
    for a message it appears to have been 'attempting' to send, then
    shuts down. If I have a 'bad' message in the queue, how would I check
    this, and erase/drop it?

    And what failure does it report? What's in qfiles/out and qfiles/retry?
    You can examine these entries with bin/dumpdb to see if there is
    something obviously bad in the recips list. You can also use the
    script at <http://www.msapiro.net/scripts/remove_recips> to list the
    recips and/or remove one or more recips.

    See the FAQ at <http://wiki.list.org/x/-IA9>.

    --
    Mark Sapiro <mark at msapiro.net> The highway is for gamblers,
    San Francisco Bay Area, California better use your sense - B. Dylan
  • Charles Gregory at Apr 24, 2009 at 8:08 pm

    On Fri, 24 Apr 2009, Mark Sapiro wrote:
    Interestingly enough, when I shutdown mailman it reports an SMTP failure
    for a message it appears to have been 'attempting' to send...
    And what failure does it report?
    Apr 24 12:26:48 2009 (7363) Low level smtp error: (4, 'Interrupted system
    call'), msgid: <BAY116-W51C78E0CDF2BD998561F7A9D740 at phx.gbl>
    Apr 24 12:26:48 2009 (7363) Low level smtp error: (104, 'Connection reset
    by peer'), msgid: <BAY116-W51C78E0CDF2BD998561F7A9D740 at phx.gbl>

    ...followed by a whole series of....
    Apr 24 12:26:48 2009 (7363) delivery to any at anydomain.tld failed with
    code -1: (104, 'Connection reset by peer')
    What's in qfiles/out and qfiles/retry?
    Where are these hiding? Tried /var/lib/mailman /usr/lib/mailman

    It's important to note that the failure logs only appear upon shutdown.
    This would suggest that mailman (and postfix) are stuck until I force
    the shutdown, which then 'resets' the connection between the two....
    See the FAQ at <http://wiki.list.org/x/-IA9>.
    No help.

    - C
  • Mark Sapiro at Apr 24, 2009 at 9:21 pm

    Charles Gregory wrote:
    On Fri, 24 Apr 2009, Mark Sapiro wrote:

    What's in qfiles/out and qfiles/retry?
    Where are these hiding? Tried /var/lib/mailman /usr/lib/mailman
    If this is a RedHat/CentOS Mailman package, there is no qfiles/
    directory per se - they ar /var/spool/mailman/out and
    /var/spool/mailman/retry. See <http://wiki.list.org/x/KYCB>
    It's important to note that the failure logs only appear upon shutdown.
    This would suggest that mailman (and postfix) are stuck until I force
    the shutdown, which then 'resets' the connection between the two....
    See the FAQ at <http://wiki.list.org/x/-IA9>.
    No help.

    Does that mean that your Python is older than 2.4 or that you are
    unwilling or unable to apply the patch to get the debugging info from
    Python's smtplib?

    I agree that Mailman and Postfix are "stuck". If your Python is 2.4.x
    or later, you can apply the patch in the FAQ to find out what is
    happening leading up to their being stuck.

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

Related Discussions

Discussion Navigation
viewthread | post
Discussion Overview
groupmailman-users @
categoriespython
postedApr 24, '09 at 4:48p
activeApr 24, '09 at 9:21p
posts6
users2
websitelist.org

2 users in discussion

Charles Gregory: 3 posts Mark Sapiro: 3 posts

People

Translate

site design / logo © 2022 Grokbase