FAQ
I am trying to run Mailman 2.1.9 on a CentOS 4 server with postfix 2.4.1. I had Mailman 2.1.5 running on a slower system and wanted to offload some apps to this newer server but am having problems. check_perms has no errors. Creating lists works fine and when postings are submitted the messages are delivered to the list admins re posting verifications.

But only a few of the messages are getting delivered. logs/smtp-failure shows many of this type of error:
"delivery to someone at gmail.com failed with code -1: (104, 'Connection reset by peer')" I am also seeing many " Low level smtp error: Server not connected, msgid" errors. I did try all the suggested fixes in the FAQ including adding the line to get error output in the logs/error file. And this file is showing successful delivery, such as
"May 26 16:33:44 2007 qrunner(20744): reply: '250 2.1.5 Ok\r\n'
May 26 16:33:44 2007 qrunner(20744): reply: retcode (250); Msg: 2.1.5 Ok
May 26 16:33:44 2007 qrunner(20744): send: 'rcpt TO:<somebody at surewest.net>\r\n""
but the majority of the messages are simply stuck in the qfiles/out and qfile/rety directory.

I am running python 2.5.1 so that I could get the error logging to work but this problem happened with python 2.3.

Any suggestions would be greatly appreciated since my mailing list no longer works at all.

Thanks.

Search Discussions

  • Stephen J. Turnbull at May 27, 2007 at 11:30 am
    Leonard Jacobs writes:
    "delivery to someone at gmail.com failed with code -1: (104,
    'Connection reset by peer')" I am also seeing many " Low level smtp
    error: Server not connected, msgid" errors.
    This seems likely to be either a Postfix problem, or maybe you've
    caught a social disease (ie, a bunch of big ISPs have decided you're a
    spammer). Either way, the first place to investigate is the Postfix
    logs.
  • Leonard Jacobs at May 28, 2007 at 2:06 pm

    ------------------------------------------------------------------------

    Today's Topics:

    1. Low level smtp error: Server not connected (Stephen J. Turnbull)

    ------------------------------------------------------------------------

    Subject:
    [Mailman-Users] Low level smtp error: Server not connected
    From:
    "Stephen J. Turnbull" <stephen at xemacs.org>
    Date:
    Sun, 27 May 2007 20:30:19 +0900
    To:
    ljacobs at shambhala.com

    To:
    ljacobs at shambhala.com
    CC:
    mailman-users at python.org

    Precedence:
    list
    MIME-Version:
    1.0
    References:
    <7c5b3fd245874985b5fa1fa9049a7b2c at shambhala.com>
    In-Reply-To:
    <7c5b3fd245874985b5fa1fa9049a7b2c at shambhala.com>
    Message-ID:
    <871wh24i7o.fsf at uwakimon.sk.tsukuba.ac.jp>
    Content-Type:
    text/plain; charset=us-ascii
    Message:
    1


    Leonard Jacobs writes:
    "delivery to someone at gmail.com failed with code -1: (104,
    'Connection reset by peer')" I am also seeing many " Low level smtp
    error: Server not connected, msgid" errors.
    This seems likely to be either a Postfix problem, or maybe you've
    caught a social disease (ie, a bunch of big ISPs have decided you're a
    spammer). Either way, the first place to investigate is the Postfix
    logs.
    Wish it were so simple. There are no obvious errors in the postfix
    maillog files. In fact postfix is receiving and sending messages fine,
    just no mailman messages are getting delivered except for the response
    to the list creator that the list was created and notifications to the
    list posters that there is a message waiting to be reviewed. Once the
    moderated poster's message is released from the queue, it triggers the
    "server not connected" errors and never actually delivers any messages.

    Strange behavior and I would appreciate any other insights and suggestions.
    ---
    [This E-mail scanned for viruses by Declude EVA]
  • Brad Knowles at May 28, 2007 at 10:57 pm

    On 5/28/07, Leonard Jacobs wrote:

    Wish it were so simple. There are no obvious errors in the postfix
    maillog files. In fact postfix is receiving and sending messages fine,
    just no mailman messages are getting delivered except for the response
    to the list creator that the list was created and notifications to the
    list posters that there is a message waiting to be reviewed. Once the
    moderated poster's message is released from the queue, it triggers the
    "server not connected" errors and never actually delivers any messages.
    Searching through the archives of the mailman-users list (in
    accordance with FAQ 1.18) turns up a number of hits. I did
    <http://www.google.com/search?q=%22Server+not+connected%22+inurl:mail.python.org/pipermail/mailman-users/>,
    but there are other searches that could also have been done.

    In the article at
    <http://mail.python.org/pipermail/mailman-users/2005-August/046366.html>,
    the problem was an inappropriate setting for SMTPPORT that needed to
    be fixed in the mm_cfg.py file.

    Then there's the article at
    <http://mail.python.org/pipermail/mailman-users/2005-September/046418.html>,
    where Mark Sapiro points out essentially the same search, and other
    issues that other people needed to resolve, in particular "setting
    SMTP_MAX_SESSIONS_PER_CONNECTION to 1 and/or setting SMTP_MAX_RCPTS
    to some small number."

    Then there's the article at
    <http://mail.python.org/pipermail/mailman-users/2004-July/038047.html>,
    and more than sixty others.

    --
    Brad Knowles <brad at shub-internet.org>, Consultant & Author
    LinkedIn Profile: <http://tinyurl.com/y8kpxu>
    Slides from Invited Talks: <http://tinyurl.com/tj6q4>

    09 F9 11 02 9D 74 E3 5B D8 41 56 C5 63 56 88 C0
  • Leonard Jacobs at May 29, 2007 at 10:21 am
    Thanks for your suggestion to read the FAQ and the associated list
    postings. I actually found those same postings and tried every
    suggestion in them that was relevant. Unfortunately nothing worked to
    resolve this annoying problem. There may be a setting that is incorrect
    and there may be something posted that explains this problem and
    provides a solution but so far I have not located anything that fixes it
    for me.

    It is strange that a pretty vanilla installation, CentOS and Postfix
    with the latest Mailman version of 2.1.9 is so hard to configure and
    use. And this after using it on a CentOS 3 box but with frequent delays
    in sending and unusal errors. So moving the list was necessary but made
    the list totally unusable. I have heard from a few others with similar
    problems and no solution, so I know I am not the only user with these
    issues.

    Any further suggestions would be appreciated.



    mailman-users-request at python.org wrote:
    Send Mailman-Users mailing list submissions to
    mailman-users at python.org

    To subscribe or unsubscribe via the World Wide Web, visit
    http://mail.python.org/mailman/listinfo/mailman-users
    or, via email, send a message with subject or body 'help' to
    mailman-users-request at python.org

    You can reach the person managing the list at
    mailman-users-owner at python.org

    When replying, please edit your Subject line so it is more specific
    than "Re: Contents of Mailman-Users digest..."

    ------------------------------------------------------------------------

    Today's Topics:

    1. Re: Low level smtp error: Server not Connected (Leonard Jacobs)
    2. Re: Low level smtp error: Server not Connected (Brad Knowles)


    ------------------------------------------------------------------------

    Subject:
    Re: [Mailman-Users] Low level smtp error: Server not Connected
    From:
    Leonard Jacobs <ljacobs at shambhala.com>
    Date:
    Mon, 28 May 2007 10:06:37 -0400
    To:
    mailman-users at python.org

    To:
    mailman-users at python.org

    Content-Transfer-Encoding:
    7bit
    Precedence:
    list
    MIME-Version:
    1.0
    Message-ID:
    <465AE1ED.5020508 at shambhala.com>
    Content-Type:
    text/plain; charset=ISO-8859-1; format=flowed
    Message:
    1



    ------------------------------------------------------------------------

    Today's Topics:

    1. Low level smtp error: Server not connected (Stephen J. Turnbull)

    ------------------------------------------------------------------------

    Subject:
    [Mailman-Users] Low level smtp error: Server not connected
    From:
    "Stephen J. Turnbull" <stephen at xemacs.org>
    Date:
    Sun, 27 May 2007 20:30:19 +0900
    To:
    ljacobs at shambhala.com

    To:
    ljacobs at shambhala.com
    CC:
    mailman-users at python.org

    Precedence:
    list
    MIME-Version:
    1.0
    References:
    <7c5b3fd245874985b5fa1fa9049a7b2c at shambhala.com>
    In-Reply-To:
    <7c5b3fd245874985b5fa1fa9049a7b2c at shambhala.com>
    Message-ID:
    <871wh24i7o.fsf at uwakimon.sk.tsukuba.ac.jp>
    Content-Type:
    text/plain; charset=us-ascii
    Message:
    1


    Leonard Jacobs writes:
    "delivery to someone at gmail.com failed with code -1: (104,
    'Connection reset by peer')" I am also seeing many " Low level smtp
    error: Server not connected, msgid" errors.
    This seems likely to be either a Postfix problem, or maybe you've
    caught a social disease (ie, a bunch of big ISPs have decided you're a
    spammer). Either way, the first place to investigate is the Postfix
    logs.
    Wish it were so simple. There are no obvious errors in the postfix
    maillog files. In fact postfix is receiving and sending messages fine,
    just no mailman messages are getting delivered except for the response
    to the list creator that the list was created and notifications to the
    list posters that there is a message waiting to be reviewed. Once the
    moderated poster's message is released from the queue, it triggers the
    "server not connected" errors and never actually delivers any messages.

    Strange behavior and I would appreciate any other insights and
    suggestions.
    ---
    [This E-mail scanned for viruses by Declude EVA]




    ------------------------------------------------------------------------

    Subject:
    Re: [Mailman-Users] Low level smtp error: Server not Connected
    From:
    Brad Knowles <brad at shub-internet.org>
    Date:
    Mon, 28 May 2007 17:57:41 -0500
    To:
    Leonard Jacobs <ljacobs at shambhala.com>, mailman-users at python.org

    To:
    Leonard Jacobs <ljacobs at shambhala.com>, mailman-users at python.org

    Precedence:
    list
    MIME-Version:
    1.0
    References:
    <465AE1ED.5020508 at shambhala.com>
    In-Reply-To:
    <465AE1ED.5020508 at shambhala.com>
    Message-ID:
    <p06240800c2810d24bbae@[192.168.1.66]>
    Content-Type:
    text/plain; charset="us-ascii" ; format="flowed"
    Message:
    2

    On 5/28/07, Leonard Jacobs wrote:

    Wish it were so simple. There are no obvious errors in the postfix
    maillog files. In fact postfix is receiving and sending messages fine,
    just no mailman messages are getting delivered except for the response
    to the list creator that the list was created and notifications to the
    list posters that there is a message waiting to be reviewed. Once the
    moderated poster's message is released from the queue, it triggers the
    "server not connected" errors and never actually delivers any messages.
    Searching through the archives of the mailman-users list (in
    accordance with FAQ 1.18) turns up a number of hits. I did
    <http://www.google.com/search?q=%22Server+not+connected%22+inurl:mail.python.org/pipermail/mailman-users/>,
    but there are other searches that could also have been done.

    In the article at
    <http://mail.python.org/pipermail/mailman-users/2005-August/046366.html>,
    the problem was an inappropriate setting for SMTPPORT that needed to
    be fixed in the mm_cfg.py file.

    Then there's the article at
    <http://mail.python.org/pipermail/mailman-users/2005-September/046418.html>,
    where Mark Sapiro points out essentially the same search, and other
    issues that other people needed to resolve, in particular "setting
    SMTP_MAX_SESSIONS_PER_CONNECTION to 1 and/or setting SMTP_MAX_RCPTS to
    some small number."

    Then there's the article at
    <http://mail.python.org/pipermail/mailman-users/2004-July/038047.html>,
    and more than sixty others.

    ------------------------------------------------------------------------

    ------------------------------------------------------
    Mailman-Users mailing list
    Mailman-Users at 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/
    ---
    [This E-mail scanned for viruses by Declude EVA]
  • Barry Warsaw at May 29, 2007 at 2:25 pm

    On May 29, 2007, at 6:21 AM, Leonard Jacobs wrote:

    Thanks for your suggestion to read the FAQ and the associated list
    postings. I actually found those same postings and tried every
    suggestion in them that was relevant. Unfortunately nothing worked to
    resolve this annoying problem. There may be a setting that is
    incorrect
    and there may be something posted that explains this problem and
    provides a solution but so far I have not located anything that
    fixes it
    for me.

    It is strange that a pretty vanilla installation, CentOS and Postfix
    with the latest Mailman version of 2.1.9 is so hard to configure and
    use. And this after using it on a CentOS 3 box but with frequent
    delays
    in sending and unusal errors. So moving the list was necessary but
    made
    the list totally unusable. I have heard from a few others with similar
    problems and no solution, so I know I am not the only user with these
    issues.

    Any further suggestions would be appreciated.
    I don't have any idea what's going on, but here are some mostly
    useless comments.

    I'm personally still on Postfix 2.3.6, and you mentioned that you're
    on Postfix 2.4.1. I wonder if there isn't some new feature, anti-
    spam defense, or other setting that is tripping up Mailman. It's
    also possible that Postfix 2.4.1 has some bug that's causing it to
    reset the connection. I don't have time to slog through the Postfix
    change logs, but I suggest you take a look to see if there's anything
    relevant.

    You might also try the Postfix 2.4.2 release candidate, and/or
    grabbing Postfix 2.3.9 to see if that improves things. IME, Postfix
    is pretty easy to build and install from source.

    If I were debugging this, my next step would probably be to write
    some sample Python applications that stress tested SMTP connections
    from a Python program. Mailman uses Python's standard smtplib
    module, so that's where I'd start. See if you can reproduce the
    problem outside of Mailman.

    I agree with previous posters that this is almost certainly a problem
    with your Postfix installation. Mailman + Postfix has been a long-
    term very stable combination, although as I said, it's possible
    there's something in Postfix 2.4 that's tripping us up now.

    Does anybody else use Postfix 2.4?

    - -Barry
  • Steve Burling at May 29, 2007 at 3:14 pm

    --On May 29, 2007 10:25:22 AM -0400 Barry Warsaw wrote:

    Does anybody else use Postfix 2.4?
    To which I reply:

    For what it's worth, my (extremely small, low-traffic) home Mailman install
    (2.1.9) uses Postfix 2.4.1 with no problems.

    --
    Steve Burling <mailto:srb at umich.edu>
    University of Michigan, ICPSR Voice: +1 734 615.3779
    330 Packard Street FAX: +1 734 647.8700
    Ann Arbor, MI 48104-2910
  • Brad Knowles at May 29, 2007 at 4:24 pm

    On 5/29/07, Barry Warsaw wrote:

    Does anybody else use Postfix 2.4?
    Actually, I'm pretty sure we are currently using it on the NTP Public
    Services Project pages at ntp.isc.org, which is the project I had
    gotten involved in years ago, prior to getting involved in Mailman.
    I got on this mailing list in order to help me figure out if it would
    be a good idea to switch from Majordomo to Mailman, and I haven't
    looked back since.

    Currently, these are the two primary open source projects where I
    spend most of my time.


    If you want to give it a test, I could install the latest postfix 2.5
    experimental or non-production patch releases on the python.org mail
    servers, so that we can give the newer code a try. ;)


    Seriously, I am quite convinced that the problem here is a
    mis-configuration between Mailman and postfix, and that neither
    program is necessarily "at fault".

    --
    Brad Knowles <brad at shub-internet.org>, Consultant & Author
    LinkedIn Profile: <http://tinyurl.com/y8kpxu>
    Slides from Invited Talks: <http://tinyurl.com/tj6q4>

    09 F9 11 02 9D 74 E3 5B D8 41 56 C5 63 56 88 C0
  • Brad Knowles at May 29, 2007 at 4:19 pm

    On 5/29/07, Leonard Jacobs wrote:

    It is strange that a pretty vanilla installation, CentOS and Postfix
    with the latest Mailman version of 2.1.9 is so hard to configure and
    use.
    Can you share with us the relevant portions of your mm_cfg.py file,
    and your postfix/main.cf file?

    Also, you did stop and restart Mailman after making these changes,
    right? Have you tried stopping and restarting it again?

    --
    Brad Knowles <brad at shub-internet.org>, Consultant & Author
    LinkedIn Profile: <http://tinyurl.com/y8kpxu>
    Slides from Invited Talks: <http://tinyurl.com/tj6q4>

    09 F9 11 02 9D 74 E3 5B D8 41 56 C5 63 56 88 C0
  • Todd Zullinger at May 30, 2007 at 3:54 am

    Brad Knowles wrote:
    On 5/29/07, Leonard Jacobs wrote:

    It is strange that a pretty vanilla installation, CentOS and
    Postfix with the latest Mailman version of 2.1.9 is so hard to
    configure and use.
    Can you share with us the relevant portions of your mm_cfg.py file,
    and your postfix/main.cf file?

    Also, you did stop and restart Mailman after making these changes,
    right? Have you tried stopping and restarting it again?
    Just a guess, but have you looked for AVC denials or other messages if
    you have SELinux enabled Leonard?

    --
    Todd OpenPGP -> KeyID: 0xBEAF0CE3 | URL: www.pobox.com/~tmz/pgp
    ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
    A diplomat is a person who can tell you to go to Hell in such a way
    that you actually look forward to the trip.
    -- Anonymous

    -------------- next part --------------
    A non-text attachment was scrubbed...
    Name: not available
    Type: application/pgp-signature
    Size: 542 bytes
    Desc: not available
    Url : http://mail.python.org/pipermail/mailman-users/attachments/20070529/5202c582/attachment-0001.pgp
  • Mark Sapiro at Jun 1, 2007 at 12:16 am

    Leonard Jacobs wrote:
    I am trying to run Mailman 2.1.9 on a CentOS 4 server with postfix 2.4.1. I had Mailman 2.1.5 running on a slower system and wanted to offload some apps to this newer server but am having problems. check_perms has no errors. Creating lists works fine and when postings are submitted the messages are delivered to the list admins re posting verifications.

    But only a few of the messages are getting delivered. logs/smtp-failure shows many of this type of error:
    "delivery to someone at gmail.com failed with code -1: (104, 'Connection reset by peer')" I am also seeing many " Low level smtp error: Server not connected, msgid" errors. I did try all the suggested fixes in the FAQ including adding the line to get error output in the logs/error file. And this file is showing successful delivery, such as
    "May 26 16:33:44 2007 qrunner(20744): reply: '250 2.1.5 Ok\r\n'
    May 26 16:33:44 2007 qrunner(20744): reply: retcode (250); Msg: 2.1.5 Ok
    May 26 16:33:44 2007 qrunner(20744): send: 'rcpt TO:<somebody at surewest.net>\r\n""
    but the majority of the messages are simply stuck in the qfiles/out and qfile/rety directory.

    Leonard,

    I don't know where you are with this. I've read the rest of the thread,
    but I don't know if you've discovered anything or not.

    Here's what I can tell you.
    From the Mailman/Python/smtplib point of view, Postfix is dropping the
    connection unexpectedly.

    You only show a small debug snippet above, but somewhere in the debug
    output there should be a point where something goes wrong in the SMTP
    transaction sequence. This should give you a clue.

    I'm guessing, since individual notices work, that postfix doesn't like
    transactions with a lot of recipients. Try setting

    SMTP_MAX_RCPTS = 1

    in mm_cfg.py and restarting mailman. If that works, you can then look
    at why Postfix doesn't like more than x recipients in a transaction.

    --
    Mark Sapiro <msapiro at value.net> The highway is for gamblers,
    San Francisco Bay Area, California better use your sense - B. Dylan
  • Leonard Jacobs at Jun 4, 2007 at 4:31 am

    On 5/29/07, Barry Warsaw wrote:

    / Does anybody else use Postfix 2.4?
    /
    Actually, I'm pretty sure we are currently using it on the NTP Public
    Services Project pages at ntp.isc.org, which is the project I had
    gotten involved in years ago, prior to getting involved in Mailman.
    I got on this mailing list in order to help me figure out if it would
    be a good idea to switch from Majordomo to Mailman, and I haven't
    looked back since.

    Currently, these are the two primary open source projects where I
    spend most of my time.


    If you want to give it a test, I could install the latest postfix 2.5
    experimental or non-production patch releases on the python.org mail
    servers, so that we can give the newer code a try. ;)


    Seriously, I am quite convinced that the problem here is a
    mis-configuration between Mailman and postfix, and that neither
    program is necessarily "at fault".

    --
    Brad Knowles <brad at shub-internet.org <http://mail.python.org/mailman/listinfo/mailman-users>>, Consultant & Author
    LinkedIn Profile: <http://tinyurl.com/y8kpxu>
    Slides from Invited Talks: <http://tinyurl.com/tj6q4>

    Still having problems with Maioman via Postfix 2.4, so I just upgraded
    to the stable patch 2.4.3. My edited mm_cfg.py file:
    MTA = 'Postfix'
    DEFAULT_URL_HOST = 'lists.shambhala.com'
    DEFAULT_EMAIL_HOST = 'lists.shambhala.com'
    add_virtualhost(DEFAULT_URL_HOST, DEFAULT_EMAIL_HOST)
    OWNERS_CAN_ENABLE_PERSONALIZATION = 1
    SMTPPORT = 25
    SMTPHOST = '192.168.1.130'

    [user]# ifconfig | grep 130
    inet addr:192.168.1.130 Bcast:192.168.1.255 Mask:255.255.255.0

    [user]# postconf | grep version
    mail_version = 2.4.3

    Yes, I have stopped & restarted Mailman after each mm_cfg.py change and
    no, there are no anti-virus or other rejections with postfix.
    Additionally I am not running SELinux on this box.

    After my recent Postfix upgrade to 2.4.3, my logs/smtp-failure contains:
    Jun 03 23:54:13 2007 (29345) Low level smtp error: (4, 'Interrupted
    system call'), msgid: <46591BE4.3030308 at shambhala.com>
    Jun 03 23:54:19 2007 (29345) Low level smtp error: (4, 'Interrupted
    system call'), msgid: <46591BE4.3030308 at shambhala.com>
    Jun 03 23:54:20 2007 (29345) Low level smtp error: (4, 'Interrupted
    system call'), msgid: <46591BE4.3030308 at shambhala.com>
    Jun 03 23:54:21 2007 (29345) Low level smtp error: (4, 'Interrupted
    system call'), msgid: <46591BE4.3030308 at shambhala.com>

    Now one clue that might be relevant is that when creating test lists and
    testing sending messages to beta users of these lists, all works fine.
    But as soon as the list contains the actual number of users, in one case
    5,100 members, in another 350 members, the mailings to the subscribers
    fail with the error as listed above. And there is no obvious throttling
    on the Postfix MTA.

    Think I should try downgrading to an earlier version of Postfix? It does
    seem like a recent upgrade of Postfix may have broken the connection
    that once did exist. Any other thoughts or suggestions would be appreciated.

    Thanks.

    ---
    [This E-mail scanned for viruses by Declude EVA]
  • Brad Knowles at Jun 4, 2007 at 4:55 am

    On 6/4/07, Leonard Jacobs wrote:

    Yes, I have stopped & restarted Mailman after each mm_cfg.py change and
    no, there are no anti-virus or other rejections with postfix.
    Additionally I am not running SELinux on this box.
    I haven't checked the earlier messages in this thread -- I'm assuming
    you have also done the stuff mentioned in FAQ 4.73 and 6.14?
    Now one clue that might be relevant is that when creating test lists and
    testing sending messages to beta users of these lists, all works fine.
    But as soon as the list contains the actual number of users, in one case
    5,100 members, in another 350 members, the mailings to the subscribers
    fail with the error as listed above. And there is no obvious throttling
    on the Postfix MTA.
    What is your "SMTP_MAX_RCPTS" setting? Check both Defaults.py and
    mm_cfg.py. Also check the value of
    "SMTP_MAX_SESSIONS_PER_CONNECTION". Then compare against their
    companion values in your postfix configuration. If you don't define
    the specific companion values in your /etc/postfix/main.cf, then you
    may need to take a look at the output of "postconf -d" and look for
    them.

    See also the postfix documentation on these values, especially
    <http://www.postfix.org/rate.html>.

    --
    Brad Knowles <brad at shub-internet.org>, Consultant & Author
    LinkedIn Profile: <http://tinyurl.com/y8kpxu>
    Slides from Invited Talks: <http://tinyurl.com/tj6q4>

    09 F9 11 02 9D 74 E3 5B D8 41 56 C5 63 56 88 C0
  • Leonard Jacobs at Jun 4, 2007 at 5:21 am

    ubject:
    Re: [Mailman-Users] Low level smtp error: Server not connected
    From:
    Leonard Jacobs <ljacobs at shambhala.com>
    Date:
    Mon, 04 Jun 2007 00:31:01 -0400
    To:
    mailman-users at python.org

    To:
    mailman-users at python.org

    Content-Transfer-Encoding:
    7bit
    Precedence:
    list
    MIME-Version:
    1.0
    References:
    <mailman.28142.1180752435.32030.mailman-users at python.org>
    In-Reply-To:
    <mailman.28142.1180752435.32030.mailman-users at python.org>
    Message-ID:
    <46639585.6060800 at shambhala.com>
    Content-Type:
    text/plain; charset=ISO-8859-1; format=flowed
    Message:
    5

    On 5/29/07, Barry Warsaw wrote:

    / Does anybody else use Postfix 2.4?
    /
    Actually, I'm pretty sure we are currently using it on the NTP Public
    Services Project pages at ntp.isc.org, which is the project I had
    gotten involved in years ago, prior to getting involved in Mailman. I
    got on this mailing list in order to help me figure out if it would
    be a good idea to switch from Majordomo to Mailman, and I haven't
    looked back since.

    Currently, these are the two primary open source projects where I
    spend most of my time.


    If you want to give it a test, I could install the latest postfix 2.5
    experimental or non-production patch releases on the python.org mail
    servers, so that we can give the newer code a try. ;)


    Seriously, I am quite convinced that the problem here is a
    mis-configuration between Mailman and postfix, and that neither
    program is necessarily "at fault".

    --
    Brad Knowles <brad at shub-internet.org
    <http://mail.python.org/mailman/listinfo/mailman-users>>, Consultant
    & Author
    LinkedIn Profile: <http://tinyurl.com/y8kpxu>
    Slides from Invited Talks: <http://tinyurl.com/tj6q4>

    Still having problems with Maioman via Postfix 2.4, so I just upgraded
    to the stable patch 2.4.3. My edited mm_cfg.py file:
    MTA = 'Postfix'
    DEFAULT_URL_HOST = 'lists.shambhala.com'
    DEFAULT_EMAIL_HOST = 'lists.shambhala.com'
    add_virtualhost(DEFAULT_URL_HOST, DEFAULT_EMAIL_HOST)
    OWNERS_CAN_ENABLE_PERSONALIZATION = 1
    SMTPPORT = 25
    SMTPHOST = '192.168.1.130'

    [user]# ifconfig | grep 130
    inet addr:192.168.1.130 Bcast:192.168.1.255 Mask:255.255.255.0

    [user]# postconf | grep version
    mail_version = 2.4.3

    Yes, I have stopped & restarted Mailman after each mm_cfg.py change
    and no, there are no anti-virus or other rejections with postfix.
    Additionally I am not running SELinux on this box.

    After my recent Postfix upgrade to 2.4.3, my logs/smtp-failure contains:
    Jun 03 23:54:13 2007 (29345) Low level smtp error: (4, 'Interrupted
    system call'), msgid: <46591BE4.3030308 at shambhala.com>
    Jun 03 23:54:19 2007 (29345) Low level smtp error: (4, 'Interrupted
    system call'), msgid: <46591BE4.3030308 at shambhala.com>
    Jun 03 23:54:20 2007 (29345) Low level smtp error: (4, 'Interrupted
    system call'), msgid: <46591BE4.3030308 at shambhala.com>
    Jun 03 23:54:21 2007 (29345) Low level smtp error: (4, 'Interrupted
    system call'), msgid: <46591BE4.3030308 at shambhala.com>

    Now one clue that might be relevant is that when creating test lists
    and testing sending messages to beta users of these lists, all works
    fine. But as soon as the list contains the actual number of users, in
    one case 5,100 members, in another 350 members, the mailings to the
    subscribers fail with the error as listed above. And there is no
    obvious throttling on the Postfix MTA.

    Think I should try downgrading to an earlier version of Postfix? It
    does seem like a recent upgrade of Postfix may have broken the
    connection that once did exist. Any other thoughts or suggestions
    would be appreciated.

    Thanks.

    ---
    [This E-mail scanned for viruses by Declude EVA]




    ------------------------------------------------------------------------

    Subject:
    Re: [Mailman-Users] Low level smtp error: Server not connected
    From:
    Brad Knowles <brad at shub-internet.org>
    Date:
    Sun, 3 Jun 2007 23:55:46 -0500
    To:
    Leonard Jacobs <ljacobs at shambhala.com>, mailman-users at python.org

    To:
    Leonard Jacobs <ljacobs at shambhala.com>, mailman-users at python.org

    Precedence:
    list
    MIME-Version:
    1.0
    References:
    <mailman.28142.1180752435.32030.mailman-users at python.org>
    <46639585.6060800 at shambhala.com>
    In-Reply-To:
    <46639585.6060800 at shambhala.com>
    Message-ID:
    <p06240803c289499ad787@[192.168.1.66]>
    Content-Type:
    text/plain; charset="us-ascii" ; format="flowed"
    Message:
    6

    On 6/4/07, Leonard Jacobs wrote:

    Yes, I have stopped & restarted Mailman after each mm_cfg.py change and
    no, there are no anti-virus or other rejections with postfix.
    Additionally I am not running SELinux on this box.
    I haven't checked the earlier messages in this thread -- I'm assuming
    you have also done the stuff mentioned in FAQ 4.73 and 6.14?
    Now one clue that might be relevant is that when creating test lists
    and
    testing sending messages to beta users of these lists, all works fine.
    But as soon as the list contains the actual number of users, in one
    case
    5,100 members, in another 350 members, the mailings to the subscribers
    fail with the error as listed above. And there is no obvious throttling
    on the Postfix MTA.
    What is your "SMTP_MAX_RCPTS" setting? Check both Defaults.py and
    mm_cfg.py. Also check the value of
    "SMTP_MAX_SESSIONS_PER_CONNECTION". Then compare against their
    companion values in your postfix configuration. If you don't define
    the specific companion values in your /etc/postfix/main.cf, then you
    may need to take a look at the output of "postconf -d" and look for them.

    See also the postfix documentation on these values, especially
    <http://www.postfix.org/rate.html>.

    ------------------------------------------------------------------------

    ------------------------------------------------------
    Mailman-Users mailing list
    Mailman-Users at 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/
    Yes, I have tried all the suggestions provided earlier in FAQ 4.73 & 6.14.
    I don't have any adjustments to the defaukt settings you mentioned in my
    mm_cfg.py file, but Defaults.py state:

    [user]# grep SESSIONS Mailman/Defaults.py
    SMTP_MAX_SESSIONS_PER_CONNECTION = 0
    [root at tao mailman]# grep RCPTS Mailman/Defaults.py
    SMTP_MAX_RCPTS = 500

    In /etc/postfix/main.cf I have the following settings:
    smtpd_client_connection_limit_exceptions = $mynetworks
    local_destination_concurrency_limit = 2
    default_destination_concurrency_limit = 100
    default_destination_recipient_limit = 50
    maximal_backoff_time = 1000s
    minimal_backoff_time = 300s

    Think these need tweaking?

    Which other postfix settings are the "companion values" to those Mailman
    ones?

    Thanks.
    ---
    [This E-mail scanned for viruses by Declude EVA]
  • Brad Knowles at Jun 4, 2007 at 5:33 am

    On 6/4/07, Leonard Jacobs wrote:

    Yes, I have tried all the suggestions provided earlier in FAQ 4.73 & 6.14.
    Could you also trim your replies? Including several levels of
    additional material from previous replies makes it very difficult to
    tell who has said what, and where the new text is.
    In /etc/postfix/main.cf I have the following settings:
    As I said, you may also need to look at the output of "postconf -d",
    because the program may be taking some defaults that have been
    compiled in as opposed to being defined in your configuration file.
    Which other postfix settings are the "companion values" to those Mailman
    ones?
    See the link I previously posted. You should pay special attention
    to the "Recipients" section further down that page (at
    <http://www.postfix.org/rate.html#recipients>).

    Just because the standard postfix source compiled-in value may be
    higher than the value you define in your mm_cfg.py file, doesn't mean
    that the people who created the binary package for your platform
    actually used that default when building the package. You need to
    check what they actually provided to you, and you can only do that
    through the output of "postconf -d".

    --
    Brad Knowles <brad at shub-internet.org>, Consultant & Author
    LinkedIn Profile: <http://tinyurl.com/y8kpxu>
    Slides from Invited Talks: <http://tinyurl.com/tj6q4>

    09 F9 11 02 9D 74 E3 5B D8 41 56 C5 63 56 88 C0
  • Mark Sapiro at Jun 5, 2007 at 1:19 am

    Leonard Jacobs wrote:
    I don't have any adjustments to the defaukt settings you mentioned in my
    mm_cfg.py file, but Defaults.py state:

    [user]# grep SESSIONS Mailman/Defaults.py
    SMTP_MAX_SESSIONS_PER_CONNECTION = 0
    [root at tao mailman]# grep RCPTS Mailman/Defaults.py
    SMTP_MAX_RCPTS = 500

    Please see my reply in this thread at
    <http://mail.python.org/pipermail/mailman-users/2007-June/057225.html>.

    In /etc/postfix/main.cf I have the following settings:
    smtpd_client_connection_limit_exceptions = $mynetworks
    local_destination_concurrency_limit = 2
    default_destination_concurrency_limit = 100
    default_destination_recipient_limit = 50

    The above seems possibly inconsistent with SMTP_MAX_RCPTS = 500

    --
    Mark Sapiro <msapiro at value.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
postedMay 26, '07 at 8:36p
activeJun 5, '07 at 1:19a
posts16
users7
websitelist.org

People

Translate

site design / logo © 2022 Grokbase