On 09-09-2014 15:23, Bruce Dubbs wrote: > Fernando de Oliveira wrote:
>> The only time I see a small difference (in the absolute, not relative, >> sense) is during shutdown, if there are many bootscripts stopping >> daemons. But this difference is so small that it is not at all a >> valuable reason for moving to systemd. I believe the time to learn, >> modify build scripts, replace them by units, debug, etc, is much longer >> than all the shutdown advantages of systemd added together. >> >> And a package designed just to make these jobs stop in parallel would be >> incomparably simpler than systemd, I believe. > > Shutdown can be sped up with a setting in rc.site. > > #Delay between TERM and KILL signals at shutdown > #KILLDELAY=3 > > Try setting to 0. Thanks, Bruce. I cannot test right now, finishing last edits after a jhalfs 7.6-rc1 and starting BLFS for tagging. But I was not complaining, I think it is shutting down quite fast no need for improvement, since you cleaned up some scripts after a previous discussion of this same theme. My point was that even if systemd is faster at shutting down, its design would not be justified as it already very good with sysV. And only for that sake, a package design would not need to become a second kernel, as systemd is becoming. -- []s, Fernando -- http://lists.linuxfromscratch.org/listinfo/lfs-dev FAQ: http://www.linuxfromscratch.org/faq/ Unsubscribe: See the above information page
