[EMAIL PROTECTED] wrote:
To clarify, the milter isn't dying... it's just refusing to accept
inputs (there's a difference). And the issue is with the milter, not
with libclamav.
In my case the milter is dying, and needs to be re-started... (started
should I say).
_______________________________________________
http://lurker.clamav.net/list/clamav-users.html
In the previous thread it was pointed out that due to the behavior of
sendmail libmilter, which is compiled either staticaly into the milter
or on some distributions as a shared object library, there is little
difference in refusing to accept connections and dying.
This is because libmilter will abort a milter that has consecutive
errors responding to a sendmail's connection attempts.
(
I was explored the possibility of changing this behavior but its not
quite clear what the absolute correct course of action is: for more
information see
http://groups-beta.google.com/group/comp.mail.sendmail/browse_frm/thread/65658b2f5797c626/8a61ae5f825ca3aa
)
Therefore milters must always do all that is neccesary to handle every
sendmail connection. If scanning is not possible they need to either
TEMPFAIL immediately or after a (very) short wait for resources to
become available or let the message pass through unscanned.
_______________________________________________
http://lurker.clamav.net/list/clamav-users.html
