Developers of Sympa

["Andreas de Pretis" <address@concealed>]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Hi,

I've a really urgent concern and my blood pressure would decrease a
bit if anyone could help/answer/advice me on this one.

We are currently using Sympa 3.4.3 and distributed a HTML-Newsletter
to 6595 subscribers yesterday evening (about 19.00/7pm). At about
01.00/1am today I found our Mailserver nearly collapsing because of
the really high load. After discovering that Sympa still distributed
the newsletter sent about 6 hours ago (the server is an Compaq Alpha
DS10 btw.) i stopped all Sympa/Sendmail processes and had to
recognize that there was something running REALLY amok within sympa
(I've attached the current configuration).

These are the facts I discovered yet:

+) within the problematic period of time (see above) there was
   only 1 processing cycle running at

   Jan 13 18:52:53 mail sympa[20405]: Processing message for news
     with priority 5, <00fa01c3d9fe$0726f4e0$b9c810ac@emmeran>

   Jan 13 18:52:53 mail sendmail[15838]: i0DHqHwn011947:
     to="| /usr/local/sympa/smrsh/queue news",
ctladdr=<address@concealed>
     (8/0), delay=00:00:36, xdelay=00:00:02, mailer=prog, pri=268393,
     dsn=2.0.0, stat=Sent

+) only 367 of 6595 have received the newsletter (discovered from
   maillog)

+) but these 367 people got up to 410 newsletter-mails (per person
!!)
   ... each one exactly like the other and the first one (no bounces,
   replies, etc.)

+) Sympa works fine since months, even years without any troubles
   for the other douzends of lists, newsletters, etc. we have defined
   on the same system

+) seconds and minutes after the distribution started there where
   several

   --
   Jan 13 18:52:58 mail sendmail[1634]: i0DHqvOM001634:
address@concealed,
     size=153785, class=-30, nrcpts=1, msgid=<news.49@anonymous>,
relay=sympa@localhost

   Jan 13 18:52:58 mail sendmail[32333]: i0DHqwwn032333:
from=<address@concealed>,
     size=154050, class=-30, nrcpts=1, msgid=<news.49@anonymous>,
proto=ESMTP, daemon=MTA,
     relay=localhost [127.0.0.1]

   Jan 13 18:53:00 mail sendmail[3962]: i0DHqwBU003962:
address@concealed,
     size=153785, class=-30, nrcpts=2, msgid=<news.49@anonymous>,
relay=sympa@localhost

   Jan 13 18:53:02 mail sendmail[11353]: i0DHr19T011353:
address@concealed,
     size=153785, class=-30, nrcpts=3, msgid=<news.49@anonymous>,
relay=sympa@localhost

   and so on ... distributing to the same addresses over and over
again:

   Jan 13 18:52:58 mail sendmail[32333]: i0DHqwwn032333:
to=<address@concealed>,
     delay=00:00:00, mailer=esmtp, pri=238050, stat=queued
   Jan 13 18:52:58 mail sendmail[1634]: i0DHqvOM001634:
address@concealed,
     address@concealed (511/511), delay=00:00:01,
xdelay=00:00:00,
     mailer=relay, pri=237785, relay=[127.0.0.1] [127.0.0.1],
dsn=2.0.0, stat=Sent (i0DHqwwn032333
     Message accepted for delivery)

   Jan 13 18:53:02 mail sendmail[24565]: i0DHr0wn024565:
to=<address@concealed>, delay=00:00:01,
     mailer=esmtp, pri=268015, stat=queued
   Jan 13 18:53:03 mail sendmail[3962]: i0DHqwBU003962:
address@concealed,address@concealed,
     address@concealed (511/511), delay=00:00:05,
xdelay=00:00:03, mailer=relay,
     pri=267785, relay=[127.0.0.1] [127.0.0.1], dsn=2.0.0, stat=Sent
(i0DHr0wn024565 Message accepted for
     delivery)

   Jan 13 18:53:03 mail sendmail[15511]: i0DHr2wn015511:
to=<address@concealed>, delay=00:00:01,
     mailer=esmtp, pri=298015, stat=queued
   Jan 13 18:53:03 mail sendmail[15511]: i0DHr2wn015511:
to=<address@concealed>, delay=00:00:01, mailer=esmtp,
     pri=298015, stat=queued
   Jan 13 18:53:03 mail sendmail[15511]: i0DHr2wn015511:
to=<address@concealed>, delay=00:00:01, mailer=esmtp,
     pri=298015, stat=queued
   Jan 13 18:53:04 mail sendmail[32079]: AUTH=client,
relay=[127.0.0.1], mech=, bits=0Jan 13 18:53:04 mail
     sendmail[11353]: i0DHr19T011353:
address@concealed,address@concealed,address@concealed,
     address@concealed (511/511), delay=00:00:03,
xdelay=00:00:02, mailer=relay, pri=297785,
     relay=[127.0.0.1] [127.0.0.1], dsn=2.0.0, stat=Sent
(i0DHr2wn015511 Message accepted for delivery)
   --

   and this goes on, and on, and on, and on ... over and over, always
the same addresses and sometimes an
   additional new one ...

   And every time sympa is invoked (not just sendmail).

+) The sympa msg queue (/var/spool/sympa/msg in our case) snapshot
I've made after discovering problems
   contains 292 msgs, 2 with newsletter-content the rest was bouncing
... (i thought bounces are only
   stored in $queuebounce?) ... in the $queuebounce directory there
are 491 bounced messages too

+) the sympa.log contains:

   --
   Jan 13 19:14:48 mail bounced[32352]: Subscriber not found in list
news : address@concealed
   Jan 13 19:14:58 mail wwsympa[7186]: [client 172.16.11.8] [user
address@concealed] [list news] do_resetbounce:
     failed deleting
/address@concealed
   Jan 13 19:14:58 mail wwsympa[7186]: [client 172.16.11.8] [user
address@concealed] [list news] do_resetbounce:
     bounces for address@concealed reset
   Jan 13 19:15:21 mail bounced[32352]: Subscriber not found in list
news : address@concealed
   Jan 13 19:20:17 mail bounced[32352]: Subscriber not found in list
news : address@concealed
   Jan 13 19:21:06 mail bounced[32352]: Subscriber not found in list
news : address@concealed
   Jan 13 19:22:22 mail bounced[32352]: Subscriber not found in list
news : address@concealed
   Jan 13 19:22:23 mail bounced[32352]: Subscriber not found in list
news : address@concealed
   Jan 13 19:22:46 mail bounced[32352]: Subscriber not found in list
news : address@concealed
   Jan 13 19:22:54 mail bounced[32352]: Subscriber not found in list
news : address@concealed
   Jan 13 19:23:15 mail bounced[32352]: Subscriber not found in list
news : address@concealed
   Jan 13 19:23:21 mail bounced[32352]: Subscriber not found in list
news : address@concealed
   Jan 13 19:23:26 mail bounced[32352]: Subscriber not found in list
news : address@concealed
   Jan 13 19:23:52 mail bounced[32352]: Subscriber not found in list
news : address@concealed
   Jan 13 19:24:14 mail last message repeated 2 times
   Jan 13 19:24:20 mail bounced[32352]: Subscriber not found in list
news : address@concealed
   Jan 13 19:24:22 mail last message repeated 3 times
   Jan 13 19:24:31 mail bounced[32352]: Subscriber not found in list
news : address@concealed
   Jan 13 19:25:01 mail last message repeated 2 times
   Jan 13 19:25:21 mail bounced[32352]: Subscriber not found in list
news : address@concealed
   Jan 13 19:25:23 mail bounced[32352]: Subscriber not found in list
news : address@concealed
   Jan 13 19:25:25 mail bounced[32352]: Subscriber not found in list
news : address@concealed
   Jan 13 19:25:52 mail bounced[32352]: Subscriber not found in list
news : address@concealed
   Jan 13 19:26:00 mail bounced[32352]: Subscriber not found in list
news : address@concealed
   Jan 13 19:26:13 mail bounced[32352]: Subscriber not found in list
news : address@concealed
   Jan 13 19:26:16 mail last message repeated 2 times
   Jan 13 19:26:36 mail bounced[32352]: Subscriber not found in list
news : address@concealed
   Jan 13 19:26:56 mail bounced[32352]: Subscriber not found in list
news : address@concealed
   Jan 13 19:27:14 mail bounced[32352]: Subscriber not found in list
news : address@concealed
   Jan 13 19:27:15 mail bounced[32352]: Subscriber not found in list
news : address@concealed
   Jan 13 19:27:42 mail bounced[32352]: Subscriber not found in list
news : address@concealed
   Jan 13 19:28:14 mail bounced[32352]: Subscriber not found in list
news : address@concealed
   Jan 13 19:28:55 mail last message repeated 3 times
   Jan 13 19:29:06 mail bounced[32352]: Subscriber not found in list
news : address@concealed
   --

   and this goes on, and on, and on till 01.00/1am ... sometimes a
new address appeared

The only two things I did with the newsletter/sympa-configuration
was:

+) adding all users (in packages to 2000 each) via the
web-interface's 'multiple add'
   feature ... (btw. as I currently saw, it seems the
'comment_subscribe'-field in the
   mysql-db contains a trailing \n or \r\n)

+) the main robot for our sympa-installation is @uptime.at ... for
this particular
   newsletter I set

        host consulting-club.cc
        anonymous_sender address@concealed (equal to the
newsletter-address; this
        is changed now, as you see in the attached configuration)

   and defined the aliases like

      address@concealed: ...
        address@concealed: ...
        ...

   instead of

        news: ...
        news-request: ...

+) resetting a bounce (see above sympa.log excerpt) while sympa still
delivered

Could anyone tell me what could have gone wrong or where I could do
some research to find out more about this really annoying incident
(some recipients which got ~400 mails where CEOs, etc. from major
austrian companies). Did anyone experience similar problems? ... Are
there any subscriber-limitations?

I'd most appreciate any kind of help or tipps ...

Thanks in advance and kind regards,
  Andreas de Pretis
  Internet-Services

  UPtime Systemlösungen GmbH
  Tel.: +43 (1) 713 61 80 - 27
  Fax : +43 (1) 713 61 80 - 10
  Mob.: +43 699 201 58 627
  Mail: <address@concealed>
  Web : <http://www.uptime.at>

-----BEGIN PGP SIGNATURE-----
Version: PGP 7.0.4

iQA/AwUBQAUz8xhaAByadA01EQLHWQCg0QA0UhGjwo/+MoRGhc/e57mtCSYAnRLJ
Y1V/F1CcjYDf/L46MDm2HQGj
=TCXA
-----END PGP SIGNATURE-----

Attachment: config
Description: Binary data