Neohapsis is currently accepting applications for employment. For more information, please visit our website www.neohapsis.com or email firstname.lastname@example.org
From: Wietse Venema (wietseporcupine.org)
Date: Thu Apr 19 2001 - 16:30:37 CDT
> On Thu, Apr 19, 2001 at 04:01:44PM -0400, Wietse Venema wrote:
> > Ben Rosengart:
> > > There are no log
> > > messages referring to 8210D9662 in between the qmgr log entries.
> > In that case, you are restarting the queue manager before it gets
> > a chance to actually deliver the message.
> Well, not intentionally ... but yes, the PIDs of the qmgrs differ.
> This is an important clue, thank you. But I can't seem to find any
> references in any of my system logs to qmgr being restarted. I
> remember earlier when it died due to too many open files, I would see
> log messages, but now I see none.
> Is it the master daemon that logs when a qmgr dies? To what facility
> and level? I want to figure out why/how qmgr is dying.
qmgr terminates voluntarily when the virtual map changes (or any
other lookup table that is used by the queue manager). In that case
the queue manager logs the following:
Apr 19 17:23:51 myhostname postfix/qmgr: table has changed -- exiting
qmgr terminates voluntarily when you type a "postfix reload" command.
In this case, the master logs the following:
Apr 19 17:23:25 myhostname postfix/master: reload configuration
In all other cases the termination of qmgr is involuntary, and
either the master daemon logs a warning that something killed qmgr,
or the qmgr screams loudly before aborting.
Postfix does not give up that easily.
To unsubscribe, send mail to majordomopostfix.org with content
(not subject): unsubscribe postfix-users