kolabmailboxfilter.php[8560]: Error: Failed to set recipient ...

Gunnar Wrobel wrobel at pardus.de
Tue Mar 20 11:29:07 CET 2007


Hi Alain,

> I got this message when a script was sending 14 emails, 2 mails failed at first try, but were finally delivered 27min later !
>
> As you can see, emails send before and after didn't encounter any problems.
> It looks like a race condition, but the server was not doing anything a that time !
> I retried to reproduce the problem, but without success !
> This is a Kolab 2.1 beta4
>
> I suspect "ockers"  to be garbage, because -1 is probably an unexpected error number !

Yes, Net/LMTP.php probably parses the response from the cyrus lmtp
socket incorrectly. I did a grep for "ockers" in the cyrus source code
and there is a section within doc/faq.html:

   Q: I'm getting messages in imapd.log like:
Sep 11 17:23:55 ogg lmtpd[773]: DBERROR db3: 16 lockers
Sep 11 17:23:55 ogg lmtpd[1409]: DBERROR db3: 17 lockers
Sep 11 17:23:56 ogg lmtpd[1508]: DBERROR db3: 9 lockers
Sep 11 17:23:56 ogg lmtpd[776]: DBERROR db3: 9 lockers
          What's wrong?
          A: Nothing is wrong. These messages are logged whenever
          Berkeley db encounters lock contention, but isn't necessarily a
          problem by themselves. This is especially likely when you have
          an empty or small duplicate delivery database and are receiving
          a large volume of e-mail.
          Berkeley db 4.0 has a bug where the number of lockers isn't
          decremented properly, causing this number to be unreliable.

So the error probably originated from the Berkeley db. But in general
it looks ok to me since the locking is necessary.

The resulting error message is not ok though. I don't know if this is
a cyrus problem (violation of the LMTP protocol) or a Net/LMTP.php
problem (parse error of the response).

Maybe somebody else knows where to submit a bug report :)

Cheers,

Gunnar

-- 
____ http://www.pardus.de _________________ http://gunnarwrobel.de _

    >> Mail at ease - Rent a kolab groupware server at p at rdus <<

p at rdus Kolab work is funded in part by KDAB and the Kolab Konsortium




More information about the users mailing list