On Tue, 7 Feb 2006 20:56:01 +0100 (CET) Juhasz Peter Karoly <[EMAIL PROTECTED]> wrote:
Replies on-list, thanks. > > I believe it happens when there are exim child processes spawned > > from Apache, and you then kill Apache. This happens e.g. if you > > have perl or php scripts that call "/usr/bin/sendmail" or similar > > to send mail, and those child processes are still running when you > > kill the webserver. > > this may be the cause, because around that time tousands of mails > were send out by a php script. What can be useful in some circumstances especially in a shared web hosting environment is to force child sites to use the -odq option to Exim (e.g. by setting sendmail_path appropriately in the case of PHP). This helps to limit the lifetime of child processes (by stopping delivery processes being forked off immediately) and prevent a misbehaving site chomping quite so many resources and filling your process table. Instead, you let Exim manage things in its own time using its queue runner mechanism. This does, of course, introduce some latency to mail deliveries from web scripts, depending on how often you start queue runners. Tim -- ## List details at http://www.exim.org/mailman/listinfo/exim-users ## Exim details at http://www.exim.org/ ## Please use the Wiki with this list - http://www.exim.org/eximwiki/
