Debian Etch, Sendmail 8.13.8 slrtscn.py (maillog parser)

[BG]

Hi,

I found that the slrtscn.py did not parse my maillog.
One example line from my maillog:
sendmail[7613]: 1db4Tccec00000001: to=<invalid@address.com>, delay=00:00:43, xdelay=00:00:00, mailer=esmtp, pri=0, relay=mail.server.com. [10.10.10.1], dsn=5.1.1, stat=User unknown

There was no regexp match
I modified the regexp to this:

Code: Select all

isstat = sre.compile ('sendmail[[0-9]+]: *([0-9A-Za-z]{6}[0-9A-Za-z]{3}[0-9A-F]{7,8}):(.*)$')

and

Code: Select all

parser = sre.compile ('^([a-z]{3} +[0-9]+ [0-9]{2}:[0-9]{2}:[0-9]{2}) +([^ ]+) +sendmail[[0-9]+]: *([0-9A-Za-z]{6}[0-9A-Za-z]{3}}[0-9A-F]{7,8}):(.*)$', sre.IGNORECASE)

after it failed in the line which contained:

Code: Select all

 mailing = int (qid[:6],16)

because the qid (1db4Tccec00000001) contains charecters not only numbers before the 6th position. So I changed it

Code: Select all

 mailing = int (qid[9:],16)

And it worked.
now I have extbounce.log file with the line:
5.1.1;0;1;0;1;stat=User unknown relay=mail.server.com. [10.10.10.1]

and I have entries in the bounce_tbl in mysql:

Code: Select all

+-----------+------------+-------------+--------+------------+---------------------+------+
| bounce_id | company_id | customer_id | detail | mailing_id | change_date         | dsn  |
+-----------+------------+-------------+--------+------------+---------------------+------+
|         1 |          1 |           1 |    511 |          1 | 2008-07-08 15:49:19 |  511 |

but unfortunately I do not have bounces in the Statictics under my mailing.
Do anybody have any idea, how can I fix this?
Thank you
bg

[BG]

Hi

is it true that the qid should contain the mailing_id?
Thank you

[ud]

The QID contains the mailing-id and the customer-id of the recipient, but only for real mailings. Test- and adminmails are not coded this way to avoid unwanted bounces for test- and adminaddresses. So the original behavior is working by design and is not a bug.

-- ud

[BG]

Hi ud,

Thank you for your answer. I'll test it with real mailings.
I think this behavior causes lot of false bug report because many of us uses test mailing to test opeemm features.
Best regards
bg.

[godber]

I found that changing the regular expression was appropriate but the suggested change to the mailing definition was incorrect and should be left as:

Code: Select all

mailing = int (qid[:6],16)

The change mentioned in the previous post only appears to be correct but will not work when you have more recipients. You will start getting "failed to map company_id" errors in your logs after the number of recipients exceeds the number of mailings.

Also as mentioned above the qid is managed by openemm and looks like:

Code: Select all

00000600000000003

or

Code: Select all

aaaaaabbbbbbbbbbb

where the 'a's are parsed out and are the mailid and the 'b's are parsed out and used as the recipient/customer id.

Austin

[Sxandy]

For next release, I hope there is a search bar for mailing page. When there are tens, hundreds of mailings, it's difficult to locate 1 specific mailing.