systemd (Was Re: tmpfs for strategic directories)
mzerqung at 0pointer.de
Tue Jun 1 02:40:25 UTC 2010
On Wed, 26.05.10 19:45, Nicolas Mailhot (nicolas.mailhot at laposte.net) wrote:
> Le dimanche 23 mai 2010 à 00:34 +0200, Lennart Poettering a écrit :
> > ATM everything looks rosy. I just finished porting over all F13
> > installed-by-default daemons to socket activation, and a few more (and
> > the patches are good enough to be upstreamable).
> For this kind of stuff I strongly suggest you do not limit yourself to
> F13 installed-by-default daemons (which are mostly well-behaved C/C++
> desktop-oriented code) but pass the reality check of converting
> important server daemons (postfix, apache, bind...) and non C/C++
> services (jboss or tomcat in java, amavisd-new or something else in
> perl, etc)
Well, I'd argue that for server software like that neither the
lazy-loading of daemons nor fast booting is really crucial. It's OK if a
server takes a bit longer to boot. It's way less important than for a
desktop or laptop. The focus of the service activation work should be on
services needed by desktop systems I guess. For everything else the
implicit activation is handy too, but not as crucial.
> Otherwise you'll replay the networkmanager drama with part of the Fedora
> users going the new laptop way, part refusing to even look at it because
> it can not translate in the server stuff they need at work, and everyone
> being very sad, unhappy, and angry at others.
Well, while network configuration is certainly very important for server
setups, super-fast booting and lazy-loading of daemons is not.
Lennart Poettering Red Hat, Inc.
lennart [at] poettering [dot] net
http://0pointer.net/lennart/ GnuPG 0x1A015CC4
More information about the devel