David Farning
The following are just some personal thoughts and notes on improving the adoption of systemd within Debian.
Reasons for adopting systemd
- There have been many advances in the Linux kernel which are effectively exposed by systemd.
- Increasing complexity of modern system require init facilities which are made available by systemd.
- A common init system across all distributions reduces work load by pushing functionality into systemd.
- A common init system across all distributions reduces work load by pushing configuration upsteam.
Concerns regarding systemd
- Distributions and developers have made large investments in code and knowledge. It will be necessary to prove the value of systemd before others are willing to throw away their previous investment and 'reinvest' in systemd.
- Distributions and developers are not going to be willing to accept regressions when adopting systemd.
- Distributions and developers are going to require a seamless fallback during transition.
Roles of systemd
- Init system to replace systemV.
- Core building block of OS.
Approach to adopting systemd in Debian
Adoption of systemd will require proving to stakeholders that systemd is better than alternative and worth the cost of conversion.
Init system
- # Drop in replacement of sysV-init. For backwards compatibility, systemd is capable of reading sysV-init scripts. Encourage gradual adoption by testers to identify bugs at rate which systemd developers can handle. # Increase coverage of unit files. Native systemd unit files are preferable to sysV-init files due to the reduced complexity and increase functionality.
Log
Feb 24 - Sent updated rsync.service to upstream as an attachment to https://bugzilla.samba.org/show_bug.cgi?id=8416
- Feb 24 - Initial edit to systemd wiki to improve clarity for new users.
# Email: <you AT SPAMFREE domain DOT org>