3 ms·

The predecessor for systemd on Fedora was upstart, not van Smoorenburg init.

by JdeBP 7y ago

mixmastamyk 7y ago
I liked upstart, but it could stand a few improvements too.
loeg 7y ago
Somewhat briefly, sure. Upstart was used in Fedora, in SysV compatibility mode, from 9 (2008-04)[0] up until 15 (2010-10)[1], or about 2.5 years. Fedora prior to 2008 used classic Redhat runlevel init (what you're calling Smoorenburg). From the first release in 2003, through Fedora 9. So we have 4.5 years of "SysV init" followed by 2.5 years of Upstart in "SysV init" compatibility mode. That said, your comment isn't especially responsive to GP; both Upstart and Systemd in Fedora continued to use a bunch of shell scripts to define startup actions. It wasn't until the systemd unit file conversion efforts[2] around 2012 that Fedora largely moved to services defined as data in "INI"-style configuration files. GP's "shell-script-based boot" is not an unfair way to characterize Fedora's boot process prior to (and including the early history of) systemd in Fedora. [0]: https://fedoraproject.org/wiki/Features/Upstart https://fedoraproject.org/wiki/Features/Upstart [1]: https://fedoraproject.org/wiki/Features/systemd https://fedoraproject.org/wiki/Features/systemd [2]: https://fedoraproject.org/wiki/Features/Systemd-unit-cleanup https://fedoraproject.org/wiki/Features/Systemd-unit-cleanup
- JdeBP 7y ago
  Miquel van Smoorenburg never worked for Redhat that I know of. It was not a "Redhat init" as you call it. Indeed, it wasn't even init. The things that we are talking about are rc, not init, in the van Smoorenburg system. Nor was there the sort of exclusive "compatibility mode" that you are envisaging in order to make your argument that mixmastamyk was referring to Upstart. Upstart could employ van Smoorenburg rc scripts, but that was not a mode. Upstart was still event-driven with job files. The van Smoorenburg rc scripts were run via what was just another job file. Were your argument valid for Upstart, it would equally be valid for systemd's mechanism for invoking van Smoorenburg rc scripts, and you would have to say that systemd is shell-script-based boot, which really makes a nonsense of mixmastamyk talking about "going back" to it from systemd. Quite clearly, going back to Upstart is not what mixmastamyk was talking about, which not least can be seen from xyr comment written, next to yours, 4 days beforehand.

The predecessor for systemd on Fedora was upstart, not van Smoorenburg init.

by JdeBP 7y ago

The predecessor for systemd on Fedora was upstart, not van Smoorenburg init.

mixmastamyk 7y ago

I liked upstart, but it could stand a few improvements too.

loeg 7y ago

Somewhat briefly, sure. Upstart was used in Fedora, in SysV compatibility mode, from 9 (2008-04)[0] up until 15 (2010-10)[1], or about 2.5 years. Fedora prior to 2008 used classic Redhat runlevel init (what you're calling Smoorenburg). From the first release in 2003, through Fedora 9. So we have 4.5 years of "SysV init" followed by 2.5 years of Upstart in "SysV init" compatibility mode. That said, your comment isn't especially responsive to GP; both Upstart and Systemd in Fedora continued to use a bunch of shell scripts to define startup actions. It wasn't until the systemd unit file conversion efforts[2] around 2012 that Fedora largely moved to services defined as data in "INI"-style configuration files. GP's "shell-script-based boot" is not an unfair way to characterize Fedora's boot process prior to (and including the early history of) systemd in Fedora. [0]: https://fedoraproject.org/wiki/Features/Upstart https://fedoraproject.org/wiki/Features/Upstart [1]: https://fedoraproject.org/wiki/Features/systemd https://fedoraproject.org/wiki/Features/systemd [2]: https://fedoraproject.org/wiki/Features/Systemd-unit-cleanup https://fedoraproject.org/wiki/Features/Systemd-unit-cleanup

JdeBP 7y ago
Miquel van Smoorenburg never worked for Redhat that I know of. It was not a "Redhat init" as you call it. Indeed, it wasn't even init. The things that we are talking about are rc, not init, in the van Smoorenburg system. Nor was there the sort of exclusive "compatibility mode" that you are envisaging in order to make your argument that mixmastamyk was referring to Upstart. Upstart could employ van Smoorenburg rc scripts, but that was not a mode. Upstart was still event-driven with job files. The van Smoorenburg rc scripts were run via what was just another job file. Were your argument valid for Upstart, it would equally be valid for systemd's mechanism for invoking van Smoorenburg rc scripts, and you would have to say that systemd is shell-script-based boot, which really makes a nonsense of mixmastamyk talking about "going back" to it from systemd. Quite clearly, going back to Upstart is not what mixmastamyk was talking about, which not least can be seen from xyr comment written, next to yours, 4 days beforehand.