Resources against systemd and alternatives

What is systemd?

systemd claims to be a good and modern replacement for SysVinit — a so called init daemon. Usually the init daemon is the first process spawned by the kernel and thus has the PID #1 and is responsible for spawning other daemons which are necessary for the OS to operate, e.g. networking, cron, syslog etc.

List of init daemons*:

- systemd

- SysVinit (home)

- OpenRC (home)

- runit (home)

- s6 (home)

- Shepherd (home)

- finit (home)

- Hummingbird (repo)

- superd (repo)

- 31init (repo)

- Dinit (repo)

- procd (repo)

- uselessd (repo; inactive)

- Upstart (home; inactive)

- InitNG (repo; inactive)

- cinit (home; inactive)

- minit (home; inactive)

- Epoch (home; inactive)

- dmaster (repo)

*_inactive_: No development activity has occurred for several years.

What is so bad about systemd?

Notable bugs and security issues:

- Datadog outage costing 5 million dollars caused by systemd upgrade

- [Widespread outage was caused on Azure, when systemd 237-3ubuntu10.54 was published to the bionic-security pocket (instances could no longer resolve DNS queries, breaking networking)](https://bugs.launchpad.net/ubuntu/+source/systemd/+bug/1988119)

- #437: timeX.google.com provide non standard time

- [#1143: PID1 getting stuck printing "systemd[1]: Time has been changed" continuously](https://github.com/systemd/systemd/issues/1143)

- #1312: restarting systemd service on dependency failure

- #1596: journalctl -r -n flags incorrectly processed

- #2402: Mount efivarfs read-only

- #2460: Showing status of service via systemctl is slow (>10s) if disk journal is used

- #4863: systemd-journald drops all bytes after '\0'

- #5644: tmpfiles: R! /dir/.* destroys root

- #6237: systemd can't handle the process privilege that belongs to user name starts with number, such as 0day

- #6369: hostnamed does not like fqdns with trailing dots

- #6381: systemd-cryptsetup: Booting with encrypted root partition fails instantly

- #6478: `systemctl` should not consider active->failed as a successful operation

- #6620: services writing to stdout become silent after journal restart

- #7184: resolved: fix loop on packets with pseudo dns types (CVE-2017-15908)

- #8579: systemd breaks my mouse

- Out-of-Bounds write in systemd-networkd dhcpv6 option handling (CVE-2018-15688)

- #6078: Applying systemd-networkd configuration disconnects all other interfaces

- System Down: A systemd-journald exploit (CVE-2018-16864, CVE-2018-16865, and CVE-2018-16866)

- #11810: Can't suspend again after suspending one time (explanation why this is bad)

- Systemd user manager interferes with ecryptfs - private directory not being unmounted (also umount of ecryptfs does not automatically clear the keyring and ecryptfs-utils: Private directory not automatically unmounted anymore on logout)

- [#2913: journald is unable to attribute messages incoming from processes that exited to their cgroup, due to /proc vs SCM_CREDS race](https://github.com/systemd/systemd/issues/2913)

Other issues arising around systemd:

- systemd will be able to kill a process if he needs/wants more RAM

- We need to enforce the use of systemd to everyone, by the creator of systemd itself.

- systemd kills background processes after user logs out (see also: [[RFE] add a way to run in a new systemd scope automatically](https://github.com/tmux/tmux/issues/428))

- Don't panic, but Linux's Systemd can be pwned via an evil DNS query

- Systemd Could Fallback to Google DNS?

- DNS search domain not removed from resolv.conf on disconnect

- systemd does not respect system wide resource limits