On Wed, 21.07.10 10:00, Adam Williamson (awilliam@xxxxxxxxxx) wrote: > > If we are breaking curmudgeony workflows it be really nice to provide > > some breadcrumbs along the way to help us old dogs learn new tricks. > > Admins are going to need to learn how to use and configure systemd > > native configs in the F14 timeframe, feedback from the "legacy" system > > tools when we need to recondition our muscle memory and are scripted > > actions would go a long way to lowering the frustration level...if the > > native/legacy tool break can't be avoided. > > It occurs to me that chkconfig isn't, in packaging terms, part of either > upstart or systemd. It's just a standalone utility, it builds from its > own SRPM. > > systemd being an open source project, its interfaces shouldn't be hard > to figure out, and hey, I seem to recall seeing quite a bit of > documentation, and Lennart seems perfectly happy to provide info when > asked. > > So...why can't any of those complaining about this just go ahead and > patch chkconfig to support systemd? It's not like Lennart can stop you, > no matter how much he twirls his moustache ;) I have no moustache. ;-) But otherwise you are right. Let me say here that the scripts in /etc/init.d resp. everything run via /sbin/service already will have the necessary glue code in there as soon as Bill merges https://bugzilla.redhat.com/show_bug.cgi?id=612728 . It will forward start/stop requests to systemctl. That means people can continue to use "/etc/init.d/foo start" and "/sbin/service foo start", and the right thing will happen and they'll learn something too. We can add similar code to chkconfig too: something that warns the admin when a native systemd unit file exists, and then redirects the command properly. It is now on my todo list, but not even near the top of the list. I would greatly appreciate if somebody would work on this, because I am currently pushing this through single-handedly everywhere else now. In fact I'd appreciate help in all areas. For example, if somebody would pick up the /var/run resp. /var/lock on tmpfs issue I'd be a happy man. Or if somebody would help me and post the unit files from http://0pointer.de/public/systemd-units/ on bugzilla, then I'd be even happier. Of course, if the other folks on this list think that bike-shedding and nitpicking is time better spent then actually doing things and helping with the integration, then they are welcome to continue wasting my time by wanting to discuss the names of the binaries we use. Jeez. Lennart -- Lennart Poettering - Red Hat, Inc. -- devel mailing list devel@xxxxxxxxxxxxxxxxxxxxxxx https://admin.fedoraproject.org/mailman/listinfo/devel
