On Thu, Feb 09, 2012 at 10:07:23PM +0100, Michael Biebl wrote: > Am 09.02.2012 17:50, schrieb Josh Triplett: > > Currently, systemd masks or overrides most of the scripts in initscripts > > in favor of its own native implementations: > ... > > However, not all of them: > > > > ~$ dpkg -L initscripts | grep '^/etc/init\.d/' | while read x ; do test -e > > "/lib/systemd/system/$(basename $x .sh).service" || echo $x; done > > /etc/init.d/rc.local > > /etc/init.d/mountoverflowtmp > > /etc/init.d/sendsigs > > /etc/init.d/umountroot > > /etc/init.d/umountfs > > /etc/init.d/skeleton > > /etc/init.d/bootlogs > > /etc/init.d/umountnfs.sh > > There is no point really in masking skeletion, sendsigs, umount* as they > are never executed by systemd.
I specifically avoided mentioning skeleton for this reason; clearly I should have filtered it out of the list entirely. :) What in systemd prevents the execution of sendsigs and umount*? I don't see those strings anywhere in systemd, and I don't see any masks for the corresponding services that systemd automatically generates for every init script. > > bootlogs does two things: update motd (which seems misplaced) and save > > Yeah, we left bootlogs in because of the motd updating. Sure. Would you consider masking or overriding it at this point, though? > > kernel messages to /var/log/dmesg. systemd's logging completely > > supersedes the need for the latter. > > Agreed. > > > That just leaves mountoverflowtmp. > > Personally I wouldn't mind masking mountoverflowtmp. It is an odd one > anyway. As sysvinit now defaults to tmpfs-for-tmp, this should probably > removed in initscripts directly though. Agreed, but those changes can happen independently. > > With all of those services masked or overriden, I think that systemd > > could then drop its dependency on the initscripts package. > > It depends on initscripts for the /run/ transition, which has all the > necessary breaks etc. In wheezy+1 we can drop that again. Oh, I see. I can understand not wanting the added complexity of having systemd also know how to transition to /run. > FWIW, I had made tests with initscripts beeing removed and it worked > just aside from insserv barfing when sysv initscripts were added and > removed. Ah, good. One more related issue: does /bin/mountpoint from initscripts need to move elsewhere, or should programs just not rely on it? (fsck.nfs can almost certainly just go away.) - Josh Triplett -- To UNSUBSCRIBE, email to [email protected] with a subject of "unsubscribe". Trouble? Contact [email protected]
