On 14.11.2011 11:10, Michal Kubeček wrote:
On Monday 14 of November 2011, Stefan Seyfried wrote:
You would need to touch each and every init script.
This was the expression used by the libcgroup developer, I assumed it to be true.
I don't think so. A simple change to /etc/init.d/rc (and /etc/init.d/boot if you want the same for boot.* scripts) would be enough.
Well, nobody did it. With systemd it's there now.
I forgot a few more benefits:
- lossless loging from the beginning
- easy display of system state (stuff like systemctl list-units --failed - do this with sysv init!) On real work systems, nobody cares what's printed on the console - simply because there is no console attached. But people care how to find out if startup worked. And they care for logs.
Again, these things either are already done for traditional init scripts or can be done easily.
Hm, unfortunately nobody has done them for 30 years. And the lossless logging will be really hard, I think. Simply because of the socket activation / buffering stuff in systemd.
(Sys V init mysteriously (and no way to find out why!) not starting single init scripts even though the S* and K* links were there is just one example).
This is a nice example of problems that are easy to debug with traditional init scripts but much harder with systemd.
No. It was actually impossible to find the cause. In the end we did something equivalent to (pseudo code) chkconfig --list > tmpfile for i in *; do chkconfig $i off; done $enable_everything_that_was enabled_before < tmpfile After that it mysteriously started working again, even though the links were the same as before. I'm not saying "we will not be seeing strange things in systemd". I'm just thinking that the debugging possibilities are at least as good in systemd than in oldskool init. -- Stefan Seyfried "Dispatch war rocket Ajax to bring back his body!" -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org