Some ideas from the list:

Small ones:

  1. systemctl show: distinguish nonexistent and existent units
  2. systemctl show: implement --sort or modify supportconfig equally
  3. systemctl: implement prefix commands in systemctl
  4. systemctl status: lists status degraded, X failed units but doesn't show which those are
  5. coredumpctl: allow deleting old cores
  6. coredump: review its performance compressing dumps
  7. systemd-analyze: allow analyzing historical boots (like journalctl -b)
  8. systemd-analyze: something as zypper versioncmp for a relation between two units
  9. jobs: display stage 1 of 3 and timeout (instead of increasing timeout just at the end of stage)

Larger implementation:

  1. use libmount in fstab-generator
  2. make used controllers configurable (and warn about unsupported directives then)
  3. support cpuset controller (in cgroup v2)

Meta:

  1. clang analyzer on systemd codebase
  2. play with/extend systemd fuzzing infrastructure

Looking for hackers with the skills:

Nothing? Add some keywords!

This project is part of:

Hack Week 17

Activity

  • over 5 years ago: dsterba liked this project.
  • over 5 years ago: mwilck liked this project.
  • over 5 years ago: hurhaj liked this project.
  • over 5 years ago: mkoutny started this project.
  • over 5 years ago: mkoutny originated this project.

  • Comments

    • mkoutny
      over 5 years ago by mkoutny | Reply

      Outcome of the hackweek:

      1. systemctl show: distinguish nonexistent and existent units -> it would be a regression, it's breaking scripts rely on LoadState=not-found

      2. systemctl show: implement --sort or modify supportconfig equally -> https://github.com/g23guy/supportutils/pull/31, also reopened https://bugzilla.suse.com/show_bug.cgi?id=1023308

      3. systemctl: implement prefix commands in systemctl -> already exists in bash-completion, effective prefices would be ambiguous

      4. systemctl status: lists status degraded, X failed units but doesn't show which those are -> systemctl list-units --failed

      5. coredumpctl: allow deleting old cores -> subject to size limitation (coredump.conf:MaxUse=) and time expiration (by default 3 days in /usr/lib/tmpfiles.d/systemd.conf), enable /usr/lib/systemd/system/systemd-tmpfiles-clean.timer

      6. coredump: review its performance compressing dumps -> https://bugzilla.suse.com/show_bug.cgi?id=1085975

      7. systemd-analyze: allow analyzing historical boots (like journalctl -b) -> the information is not mined from journal but from individual unit properties, i.e. we only have enough information for the last boot :-(

      8. systemd-analyze: something as zypper versioncmp for a relation between two units ->https://github.com/werkov/systemd/tree/analyze-cmp

      I run out of time for the rest of the tasks. Besides that I also prepared a mkosi config for openSUSE.

    Similar Projects

    This project is one of its kind!