On 10/08/2013 11:24 AM, Eric Shubert wrote:
On 10/08/2013 11:00 AM, Brent Gardner wrote:
On my toasters, qtp-sa-update executes every day at 4:02am.  This stops
then restarts spamassassin.

Will this interfere with delivery of mail?

A message was sent from a gmail address to an address on one of my
toasters this morning.

The sender got a bounce: 550 5.1.1 sorry, no mailbox here by that name
(chkuser)

Checking chkuser entries in the toaster's smtp log
(/var/log/qmail/smtp/current), I see that chkuser created the reject log
entry at 4:02am.

The mailbox definitely exists and is used regularly.

Running Centos 5.9 32-bit

Toaster packages:

     autorespond-toaster-2.0.5-1.4.0
     clamav-toaster-0.97.8-1.4.4
     control-panel-toaster-0.5-1.4.0
     courier-authlib-toaster-0.59.2-1.3.10
     courier-imap-toaster-4.1.2-1.3.10
     daemontools-toaster-0.76-1.3.6
     ezmlm-cgi-toaster-0.53.324-1.3.6
     ezmlm-toaster-0.53.324-1.3.6
     isoqlog-toaster-2.1-1.3.7
     libdomainkeys-toaster-0.68-1.3.6
     libsrs2-toaster-1.0.18-1.3.6
     maildrop-toaster-2.0.3-1.3.8
     maildrop-toaster-devel-2.0.3-1.3.8
     qmailadmin-toaster-1.2.16-1.4.1
     qmailmrtg-toaster-4.2-1.3.7
     qmail-pop3d-toaster-1.03-1.3.22
     qmail-toaster-1.03-1.3.22
     qmailtoaster-plus-0.3.2-1.4.19
     qmailtoaster-plus.repo-0.2-2
     ripmime-toaster-1.4.0.6-1.3.6
     send-emails-toaster-0.5-1.4.0
     simscan-toaster-1.4.0-1.4.0
     spamassassin-toaster-3.3.2-1.4.3
     squirrelmail-toaster-1.4.22-1.4.0
     ucspi-tcp-toaster-0.88-1.3.9
     vpopmail-toaster-5.4.33-1.4.0
     vqadmin-toaster-2.3.7-1.4.1


Regards,


Brent Gardner





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

I don't know for sure. I would expect qmail to return a 4xx (temporary) failure message if spamassassin (via simscan) wasn't available. At least that's what *should* happen.

The times do appear to be more than coincidental though. Can you run a test? Shut down spamd, send a message to QMT, and see what happens.


To test, I changed firewall rules to block inbound from the public, then as root ran these commands copied from qtp-sa-update:

    svc -d /var/qmail/supervise/spamd /var/qmail/supervise/spamd/log
    svc -t /var/qmail/supervise/spamd /var/qmail/supervise/spamd/log


qmailctl-stat then showed:

    authlib: up (pid 12661) 34242 seconds
    clamd: up (pid 11199) 254 seconds
    imap4: up (pid 12646) 34242 seconds
    imap4-ssl: up (pid 12652) 34242 seconds
    pop3: up (pid 12663) 34242 seconds
    pop3-ssl: up (pid 12653) 34242 seconds
    send: up (pid 12659) 34242 seconds
    smtp: up (pid 12657) 34242 seconds
    spamd: down 14 seconds, normally up
    submission: up (pid 12655) 34242 seconds
    authlib/log: up (pid 12647) 34242 seconds
    clamd/log: up (pid 11200) 254 seconds
    imap4/log: up (pid 12650) 34242 seconds
    imap4-ssl/log: up (pid 12651) 34242 seconds
    pop3/log: up (pid 12662) 34242 seconds
    pop3-ssl/log: up (pid 12654) 34242 seconds
    send/log: up (pid 12660) 34242 seconds
    smtp/log: up (pid 12658) 34242 seconds
    spamd/log: down 14 seconds, normally up
    submission/log: up (pid 12656) 34242 seconds


I used blat on a Windows box to send a message to the machine. It went through without trouble, here's the output from blat:

    <<<getline<<< 220 mx01.example.com - THE QUENCHIEST ESMTP
    >>>putline>>> EHLO my-windows-box.example.local
    <<<getline<<< 250-mx01.example.com - THE QUENCHIEST
    <<<getline<<< 250-XXXXXXXA
    <<<getline<<< 250-PIPELINING
    <<<getline<<< 250-8BITMIME
    <<<getline<<< 250-SIZE 26214400
    <<<getline<<< 250 XXXXXXXXB
    Sending test-mx01.txt to bgard...@example.com
    Subject: Test mail via mx01.example.com
    Login name is itadm...@example.com
    >>>putline>>> MAIL FROM:<itadm...@example.com>
    <<<getline<<< 250 ok
    >>>putline>>> RCPT TO:<bgard...@example.com>
    <<<getline<<< 250 ok
    >>>putline>>> DATA
    <<<getline<<< 354 go ahead
    <<<getline<<< 250 ok 1381257125 qp 17696
    >>>putline>>> QUIT
    <<<getline<<< 221 mx01.example.com - THE QUENCHIEST


So I guess my theory was wrong and I've got some more digging to do ;)


Thanks,

Brent Gardner



---------------------------------------------------------------------
To unsubscribe, e-mail: qmailtoaster-list-unsubscr...@qmailtoaster.com
For additional commands, e-mail: qmailtoaster-list-h...@qmailtoaster.com

Reply via email to