On Thu, 14 Jul 2005, Dr Robert Young wrote:
> I am trying to id the cause of a milter-spamc-> spamd timeout. There
> are all the lines from the maillog. Is it possible to tell from the
> following 'where' (ie which test or even type of test) it might have
> timed out on? The spam passed on through without ever being seem by
> spamd (no X-Spam header in the email), which is the default I believe
> if the INPUT_FILTER argument on sendmail has no "F=" clause? .
>
> [EMAIL PROTECTED] /root]# cat /var/log/maillog | grep j6DKkGNJ020346
> Jul 13 16:46:16 email1 sendmail[20346]: j6DKkGNJ020346: Milter
> (milter-spamc): init success to negotiate
> Jul 13 16:46:16 email1 sendmail[20346]: j6DKkGNJ020346: Milter: connect
> to filters
[snip..]
> Jul 13 16:46:17 email1 sendmail[20346]: j6DKkGNJ020346:
> milter=milter-spamc, action=body, continue
> Jul 13 16:48:17 email1 milter-spamc[20347]: 47567 j6DKkGNJ020346:
> timeout before input from SPAMD server
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
> Jul 13 16:48:17 email1 milter-spamc[20347]: 47567 j6DKkGNJ020346: SPAMD
> status line failure
> Jul 13 16:48:17 email1 sendmail[20346]: j6DKkGNJ020346: Milter accept:
> message
There's three different processes involved in this scenario, sendmail,
milter-spamc and spamd, thus two different sets of IPC conversations:
sendmail <-> milter-spamc and milter-spamc <-> spamd
The stuff in the sendmail.cf file ("F=") only applies to the conversation
between sendmail & milter-spamc. That process does not have any problems,
it's the conversations between milter-spamc & spamd that is having
trouble. (see '^^^' underlined part in above logfile excerpt).
So you need to look at the milter-spamc <-> spamd communications to see
what's wrong. (IE is spamd running, is it blocked by some kind of
ip-filter rule, does milter-spamc have the correct address for the
spamd socket, etc).
First test spamd by feeding it a message using spamc. If that does
work then look closely at how milter-spamc is trying to 'talk' to spamd.
Try turning up the logging level on milter-spamc, the sendmail
logging won't help you any as the problem isn't at the sendmail end.
Dave
--
Dave Funk University of Iowa
<dbfunk (at) engineering.uiowa.edu> College of Engineering
319/335-5751 FAX: 319/384-0549 1256 Seamans Center
Sys_admin/Postmaster/cell_admin Iowa City, IA 52242-1527
#include <std_disclaimer.h>
Better is not better, 'standard' is better. B{
