Re: Code burn-in: postscreen/verify cache cleanup
- * Wietse Venema <wietse@...>:
> Systems that run close to the capacity limit probably should notI deployed 20091230-nonprod before I went to town this evening and
> expire caches but simply rotate them. I already have a version of
> Postfix that allows you to turn off cache cleanup.
until now, there is not a single warning emitted by postscreen in my
logs, although the server has been handling about twice the mail
volume and about three times as many connections during this time
period as in the previous days. I really don't know why. Sorry.
As for the Debian packaging, I sorted out the problem with dynamic map
loading, replaced a few instances of $config_directory/postfix-script
in conf/postfix-script to use $daemon_directory/postfix-script
instead, added the creation of a symlink to the main dynamicmaps.cf
file when creating a new instance in conf/postmulti-script, made the
chroot preparation in Debian's init.d script aware of multiple
instances and modified the packaging scripts so that a "purge"
operation will now remove config and queue directories of all
instances. Before bothering the Debian developer, I'd need a few
people willing to test those changes - please contact me off-list.