Yes I work as a Unix admin, and I’m getting so sick of “systemd”, that complete ball of crap that Poettering has foisted upon the Linux community. His poor quality of work has singlehandedly caused the most problems I’ve had to face in my day-to-day work in the past year, through his shoddy code, his abysmal system design and his complete lack of knowledge on how to do things the Linux way.
Things that systemd breaks, in no particular order, and I’ll update this as I see them:
- The use of binary log files breaks the simplicity of administration with text files.
- systemd returns silent or incorrect return status from script starts. Nothing quite like a return status of “success” when the script actually failed.
- Attempting to deprecate /etc/fstab by using its own mounting system, breaking a known-working system
- Because systemd is so large and broken, it no longer has a small target area for any attacks on PID1. This is a major security issue.
- systemd breaks the philosophy of “do one job, and do it well”, as it subscribes the to the known-broken monolithic philosophy as seen in Microsoft products, as it attempts to be a jack-of-all-trades and a master of none.
- Poettering is also a poor responder to either bug reports or any warranted criticism of his project. There are many real bugs marked as wont-fix on the systemd bug tracker, and his comments on bugs are unprofessional and suggest that he has some issues to deal with, that should have no place in someone that manages a core server project.
I’ve seen systemd machines fail to reboot because systemd silently crashes on shutdown, with no console errors and no errors logged. That’s been fun to troubleshoot.
In short, systemd is utterly crap and is breaking Linux. There’s no surprise that there’s a groundswell of Linux admins that are working to remove systemd from their Linux distribution of choice. I choose to use Devuan, which is Debian that has been repaired by the removal of systemd.
See here for a great analysis of the situation.