index.md (20465B)
1 systemd is the best example of Suck. 2 ==================================== 3 There is a menace which is spreading like a disease throughout the Linux 4 world, it is called [systemd](https://github.com/systemd/systemd). 5 6 Short Historical Summary 7 ------------------------ 8 systemd is a replacement for the standard init command, which normally runs 9 as process id 1 on initialisation of a UNIX bootup. There has been a movement, 10 especially around the [Red Hat](http://www.redhat.com)-related developers to 11 copy [Microsoft Windows](http://www.microsoft.com/) and all of its features. 12 Now this interpretation of how a userspace should look like is implemented and 13 was introduced with big criticism and change in the Open Source world into 14 many distributions. The [debacle in Debian](https://wiki.debian.org/Debate/initsystem/systemd) 15 is the best example in how to not introduce such a changing technology into a 16 distribution. 17 18 What PID 1 Should Do 19 ------------------------ 20 When your system boots up the kernel is executing a given binary in its known 21 namespace. To 22 see what are the only tasks the application running as pid 1 has to do, see 23 [sinit](//git.suckless.org/sinit/). Just wait for child process to reap 24 and run some other init scripts. 25 26 The systemd Chronicles 27 ---------------------- 28 Now follows a collection of links related to all the features systemd tries to 29 enforce upon you as a Linux user, because »they know better«. 30 31 Please add all the links you can find! Generations after us have to know how 32 to not do software and system development! See the 33 [wiki instructions](//suckless.org/wiki) on how to contribute to this 34 page. 35 36 The Never Ending Part 37 --------------------- 38 * Your link here. 39 40 * [The SystemD Platform: an Unbiased Critique](http://soggywizards.com/tips/code/os/initialization/systemd/good-bad-ugly/) 41 42 Part 4 43 ------ 44 * [systemd breaks POSIX IPC](https://github.com/systemd/systemd/issues/2039) 45 * [systemd-logind deletes your message queues](https://knzl.de/systemd-removeipc) 46 * [systemd breaks the mouse(again)](https://github.com/systemd/systemd/issues/8579) 47 48 Part 3 49 ------ 50 [NEWS revision 76153ad45f09b6ae45464f2e03d3afefbb4b2afe](https://github.com/systemd/systemd/blob/76153ad45f09b6ae45464f2e03d3afefbb4b2afe/NEWS) 51 52 * [systemd does {,U}EFI bootload](https://github.com/systemd/systemd/blob/76153ad45f09b6ae45464f2e03d3afefbb4b2afe/NEWS#L1306) 53 * Should systemd's PID be changed from 1 to a negative, or imaginary, 54 number? It now exists before the kernel itself, during a bootup. 55 See also [systemd-boot](https://wiki.archlinux.org/index.php/systemd-boot). 56 * [systemd replaces sudo and su](https://github.com/systemd/systemd/blob/76153ad45f09b6ae45464f2e03d3afefbb4b2afe/NEWS#L1233) 57 * Please note the command name, machinectl and 58 its [features at the manpage](https://www.freedesktop.org/software/systemd/man/machinectl.html). 59 In exchange for a program which contains sudo, su and kill 60 (and does some functions which historically ssh/telnet did), bare metal 61 users have a tons of bloat and a lot of things to disable, if even 62 possible, useful only to people which deal with virtual machines. 63 * [systemd-journald can do log-rotate](https://github.com/systemd/systemd/blob/76153ad45f09b6ae45464f2e03d3afefbb4b2afe/NEWS#L1073) 64 * Being journal files binaries written with easily corruptable 65 transactions, does this feature make the log unreadable at times? 66 * [Transient units](https://github.com/systemd/systemd/blob/76153ad45f09b6ae45464f2e03d3afefbb4b2afe/NEWS#L708) 67 * Temporary services, because we love to reinvent procps, forking, 68 nohup and lsof. 69 * [systemd does socat/netcat](https://github.com/systemd/systemd/blob/76153ad45f09b6ae45464f2e03d3afefbb4b2afe/NEWS#L546) 70 * [systemd-logind does sighup and nohup](https://github.com/systemd/systemd/blob/76153ad45f09b6ae45464f2e03d3afefbb4b2afe/NEWS#L274) 71 * Logout is equivalent to shutting off the machine, so you will NOT 72 have any running program after logout, unless you inform your init system. 73 * [systemd-nspawn can patch at will any kind of file in a container](https://github.com/systemd/systemd/blob/76153ad45f09b6ae45464f2e03d3afefbb4b2afe/NEWS#L391) 74 * Paired with transient units and user escalation performable remotely, 75 this can mean that if you house VPS instances somewhere, your 76 hosting provider has means and tools to spy, modify, delete any 77 kind of content you store there. Encrypt everything, read your TOS. 78 * [systemd does UNIX nice](https://github.com/systemd/systemd/blob/76153ad45f09b6ae45464f2e03d3afefbb4b2afe/NEWS#L425) 79 80 from 2016-08-06 81 82 Part 2 83 ------ 84 * [systemd locks down /etc and makes it read-only](http://bugzilla.redhat.com/show_bug.cgi?id=1350450) 85 * [systemd now does your DNS](https://lists.dns-oarc.net/pipermail/dns-operations/2016-June/014964.html) 86 * [systemd hates when you adapt your system (graphics on other than vt1)](https://bugzilla.redhat.com/show_bug.cgi?id=817186) 87 88 Part 1 89 ------ 90 (If you are adding more parts, link to the right revision and not just the 91 plain file.) 92 93 [NEWS revision 3f80af378331748e1373d4c189a8e20cddb735c9](http://cgit.freedesktop.org/systemd/systemd/tree/NEWS?id=2d1ca11270e66777c90a449096203afebc37ec9c) 94 95 * [logind should wait](http://cgit.freedesktop.org/systemd/systemd/tree/NEWS?id=2d1ca11270e66777c90a449096203afebc37ec9c#n2089) 96 * systemd was introduced to decrease the boot up time. Now that they 97 do not understand all implications and dependencies, let us add some 98 artifical time we found out might work for the developers 99 laptops. More on this small world hypothesis of the systemd 100 developers below. 101 * [screen brightness](http://cgit.freedesktop.org/systemd/systemd/tree/NEWS?id=2d1ca11270e66777c90a449096203afebc37ec9c#n1852) 102 * Screen brightness is something that should crash your boot up when 103 it is not working. 104 * [hostnamed](http://cgit.freedesktop.org/systemd/systemd/tree/NEWS?id=2d1ca11270e66777c90a449096203afebc37ec9c#n1763) 105 * There really should be a process running which exposes the content 106 of a file. Complexity is without cost. 107 * [seqnum removed](http://cgit.freedesktop.org/systemd/systemd/tree/NEWS?id=2d1ca11270e66777c90a449096203afebc37ec9c#n1742) 108 * The sequential ordering of requests was one reason why udevd was 109 introduced. Now remove it, because the developer laptops do not have 110 a problem anymore. 111 * [floppy group removed](http://cgit.freedesktop.org/systemd/systemd/tree/NEWS?id=2d1ca11270e66777c90a449096203afebc37ec9c#n1703) 112 * Because we know what is right to know about groups. This is just one 113 example of the mass of group name dependencies systemd is adding. 114 See sinit for how to not need such dependencies. 115 * [sysv removed](http://cgit.freedesktop.org/systemd/systemd/tree/NEWS?id=2d1ca11270e66777c90a449096203afebc37ec9c#n1651) 116 * We have won. Now remove all remains of our defeated enemy as 117 fast as we can. As said in the beginning of the systemd crusade 118 against the UNIX infidels: »You can patch it out.« It is no more 119 there. 120 * [abnormal processes](http://cgit.freedesktop.org/systemd/systemd/tree/NEWS?id=2d1ca11270e66777c90a449096203afebc37ec9c#n1627) 121 * Now systemd is getting deep into philosophy. What is »abnormal«? 122 Well, let us just define it. There is no technical merit to accept 123 this. 124 * [systemd-resolved](http://cgit.freedesktop.org/systemd/systemd/tree/NEWS?id=2d1ca11270e66777c90a449096203afebc37ec9c#n1578) 125 * Every configuration file needs its own process and service. 126 * Symlinks are a good way to solve all world problems. 127 * [new is better](http://cgit.freedesktop.org/systemd/systemd/tree/NEWS?id=2d1ca11270e66777c90a449096203afebc37ec9c#n1571) 128 * The systemd development process is flawed by always assuming »new is 129 best«. 130 * Network configuration should be in my init process. 131 * [remote pid 1](http://cgit.freedesktop.org/systemd/systemd/tree/NEWS?id=2d1ca11270e66777c90a449096203afebc37ec9c#n1500) 132 * »Everything will end up having a remote API.« I wonder when systemd 133 will understand MIME and e-mail. 134 * [init does man](http://cgit.freedesktop.org/systemd/systemd/tree/NEWS?id=2d1ca11270e66777c90a449096203afebc37ec9c#n1465) 135 * My init process is too big, it needs its own file hierarchy and an 136 abstraction layer to find paths. 137 * [factory reset](http://cgit.freedesktop.org/systemd/systemd/tree/NEWS?id=2d1ca11270e66777c90a449096203afebc37ec9c#n1401) 138 * Welcome to the Windows OEM world: Factory reset for Linux! Of course 139 it is in your init process. 140 * [system runs runs](http://cgit.freedesktop.org/systemd/systemd/tree/NEWS?id=2d1ca11270e66777c90a449096203afebc37ec9c#n1397) 141 * Exactly. The predisposition of being able to call such a complex 142 command does not imply the running system. Let's check it again. 143 * [clean up directories](http://cgit.freedesktop.org/systemd/systemd/tree/NEWS?id=2d1ca11270e66777c90a449096203afebc37ec9c#n1306) 144 * There is another monster in systemd, it handles tmp files. 145 There are just some cases before it was introduced to have to clean up 146 a directory in the file tree. Now there are hundreds. And easily 147 another case can be added! Of course your init process does that. 148 * [firstboot](http://cgit.freedesktop.org/systemd/systemd/tree/NEWS?id=2d1ca11270e66777c90a449096203afebc37ec9c#n1259) 149 * »Interactive queries« pulls in many dependencies. Let us have it in 150 every installation out there on by default. Of course in pid 1. 151 * [journald](http://cgit.freedesktop.org/systemd/systemd/tree/NEWS?id=2d1ca11270e66777c90a449096203afebc37ec9c#n1234) 152 * This is a bit longer. Because the systemd developers were not able 153 to contribute to any syslog daemon they had to write their own with 154 some binary format so the principle of being able to read your log 155 files after a critical crash was violated. To be nice invaders the old 156 systems were allowed to order the system log through the specified 157 mechanism. Now that they implemented our specifics, turn off the 158 neutral syslog delivery. You will see this pattern of »now that we 159 conquered your culture, obey« more often in systemd. 160 * [systemd-terminal](http://cgit.freedesktop.org/systemd/systemd/tree/NEWS?id=2d1ca11270e66777c90a449096203afebc37ec9c#n1224) 161 * Why does the kernel have tty handling? So in serious situations you 162 will be able to debug it over the last standing PIN on your 163 motherboard. Let us remove this, run it in pid 1. 164 * [networkd is your oppressor](http://cgit.freedesktop.org/systemd/systemd/tree/NEWS?id=2d1ca11270e66777c90a449096203afebc37ec9c#n1216) 165 * Premature optimisation of IP configurations always leads to misery. 166 * [We do not understand broadcast](http://cgit.freedesktop.org/systemd/systemd/tree/NEWS?id=2d1ca11270e66777c90a449096203afebc37ec9c#n1211) 167 * With the growth of systemd in complexity and the new depending 168 software the implications of the added hacks are increasing. 169 * [Timezone hack](http://cgit.freedesktop.org/systemd/systemd/tree/NEWS?id=2d1ca11270e66777c90a449096203afebc37ec9c#n1193) 170 * systemd is too complex for such a simple transaction with the 171 kernel. Do not inform the kernel and add another assumption which is 172 only documented in the changelog. 173 * [the web is a session](http://cgit.freedesktop.org/systemd/systemd/tree/NEWS?id=2d1ca11270e66777c90a449096203afebc37ec9c#n1168) 174 * For the future generations: In 2015 nearly everyone was trying to 175 connect all you do to the web. In the UNIX model it is just a client 176 which should be simple and integrate into the system. systemd is 177 doing it the other way around and does integrate itself into the 178 web. 179 * The discussion about why my pid 1 is handling sessions is discussed 180 in other points. 181 * [pid 1 does DNS](http://cgit.freedesktop.org/systemd/systemd/tree/NEWS?id=2d1ca11270e66777c90a449096203afebc37ec9c#n1128) 182 * This change adds another open door to your pid 1, adds caches, new 183 APIs and of course it will never fail and break systemd on your 184 initial boot. 185 * They are planning (2015-07) to add multicast DNS to your pid 1. 186 Multicast DNS implies an enforced timeout to know if someone answered. 187 systemd was about speed, you know. Look at avahi in how to ruin this 188 concept for Linux. It has the same developer flaws as systemd: Too 189 many forced dependencies, you need to implement its API and when it is 190 taking over you are doomed. 191 * [policykit](http://cgit.freedesktop.org/systemd/systemd/tree/NEWS?id=2d1ca11270e66777c90a449096203afebc37ec9c#n1113) 192 * There is a bus inside your pid 1 and if it crashes you are flawed. 193 Exactly, there are a thousand more cases of errors that could occur 194 and make your system unbootable. Instead of using a separation of 195 functions, add everything to a big bus. 196 * Of course when you are using a misdesign like dbus you need to add 197 interactions over the bus to add features you forgot in the initial 198 design. Now let us have our pid 1 have to query for the permission 199 to boot. 200 * »Interactive authentication« will produce you a graphical input form 201 on your serial line. 202 * [Calendar](http://cgit.freedesktop.org/systemd/systemd/tree/NEWS?id=2d1ca11270e66777c90a449096203afebc37ec9c#n1044) 203 * As you see, your pid 1 should handle your calendars and cron jobs 204 too. 205 * [utmp should go](http://cgit.freedesktop.org/systemd/systemd/tree/NEWS?id=2d1ca11270e66777c90a449096203afebc37ec9c#n994) 206 * We have taken over your culture, now die! Another flaw in the 207 systemd small world theory: When something is getting optional it 208 will be removed. 209 * [password agent](http://cgit.freedesktop.org/systemd/systemd/tree/NEWS?id=2d1ca11270e66777c90a449096203afebc37ec9c#n964) 210 * »Interactive authentication« 211 * [udev timeout](http://cgit.freedesktop.org/systemd/systemd/tree/NEWS?id=2d1ca11270e66777c90a449096203afebc37ec9c#n944) 212 * Instead of patching the kernel to add a simple solution, add a hack. 213 Only the systemd developers tell you when it is allowed to wait or 214 sleep in userspace. The rest obey our orders! 215 * [systemd-pm](http://cgit.freedesktop.org/systemd/systemd/tree/NEWS?id=2d1ca11270e66777c90a449096203afebc37ec9c#n924) 216 * Power management is required on boot up. 217 * [user systemd units](http://cgit.freedesktop.org/systemd/systemd/tree/NEWS?id=2d1ca11270e66777c90a449096203afebc37ec9c#n898) 218 * What can go wrong when you are adding more paths that are read, 219 parsed and executed? 220 * [hack the reload](http://cgit.freedesktop.org/systemd/systemd/tree/NEWS?id=2d1ca11270e66777c90a449096203afebc37ec9c#n888) 221 * First systemd was adding »better features« like socket activation to 222 make developers use their mechanism for daemons. They hit the 223 proprietary wall of disgust with this changelog entry. systemd is 224 too big and you will lose your face if you change the misdesign. Now 225 add another hack because we can do it. Big empires fall too and 226 sadly have too many casualties when they are falling. :( 227 * [X11 in systemd](http://cgit.freedesktop.org/systemd/systemd/tree/NEWS?id=2d1ca11270e66777c90a449096203afebc37ec9c#n783) 228 * Of course graphics were missing in pid 1. 229 * [complexity is purity](http://cgit.freedesktop.org/systemd/systemd/tree/NEWS?id=2d1ca11270e66777c90a449096203afebc37ec9c#n738) 230 * You will of course need PPPoE when you do parallel bootup. Every 231 1000 lines of code add one critical bug you never fix. 232 * [gateway hostname](http://cgit.freedesktop.org/systemd/systemd/tree/NEWS?id=2d1ca11270e66777c90a449096203afebc37ec9c#n700) 233 * We rule the world so we are above IETF and IANA. Now add our own 234 hostnames that of course won't add another assumption. 235 * [no editor in systemd](http://cgit.freedesktop.org/systemd/systemd/tree/NEWS?id=2d1ca11270e66777c90a449096203afebc37ec9c#n689) 236 * This one is a setback. Why is there no default editor in systemd in 237 case of factory reset? 238 * [8x ctrl + alt + del](http://cgit.freedesktop.org/systemd/systemd/tree/NEWS?id=2d1ca11270e66777c90a449096203afebc37ec9c#n610) 239 * In systemd you press eight times Ctrl+Alt+Del to trigger reboot. 240 * [privacy policy](http://cgit.freedesktop.org/systemd/systemd/tree/NEWS?id=2d1ca11270e66777c90a449096203afebc37ec9c#n594) 241 * For the next generations: In 2015 privacy was a big issue because of 242 the mentioned hard-wiring between the web and software. 243 As you can see, every commit which adds some preparation for a 244 feature adds another intepretation of what will be a major 245 assumption in a next release. If you 246 handle privacy you will have some features depending on that user 247 decision and of course the factory reset default value. 248 * Why didn't they use XML for /etc/os-release? 249 * [fds cache](http://cgit.freedesktop.org/systemd/systemd/tree/NEWS?id=2d1ca11270e66777c90a449096203afebc37ec9c#n502) 250 * We have talked about misdesign, too-big-to-fail and world 251 domination. This is the next example of a hack that is prone to 252 fail. 253 * [umount -rf](http://cgit.freedesktop.org/systemd/systemd/tree/NEWS?id=2d1ca11270e66777c90a449096203afebc37ec9c#n380) 254 * This is umount for dummies. Just do one thing – right. 255 * [libudev will be orphaned](http://cgit.freedesktop.org/systemd/systemd/tree/NEWS?id=2d1ca11270e66777c90a449096203afebc37ec9c#n291) 256 * With the advent of udevd there was a compatibility to its complexity 257 called libudev. X11 uses it to query the changing devices. And of 258 course make it a non-independent API in systemd. Why? You can guess 259 why: Defeating the infidels. 260 * [fsck indirections](http://cgit.freedesktop.org/systemd/systemd/tree/NEWS?id=2d1ca11270e66777c90a449096203afebc37ec9c#n281) 261 * When there was syslog fsck did output errors to your display as 262 easy as possible. Now add a hack to have this possible again. 263 * [systemd-importd](http://cgit.freedesktop.org/systemd/systemd/tree/NEWS?id=2d1ca11270e66777c90a449096203afebc37ec9c#n236) 264 * This is pure evil. Your pid 1 is now able to import complete system 265 images over the network and show them to you as your running system. 266 There is nothing that can go wrong. 267 * [CGI for systemd](http://cgit.freedesktop.org/systemd/systemd/tree/NEWS?id=2d1ca11270e66777c90a449096203afebc37ec9c#n166) 268 * The web thing has been discussed before. 269 * [kdbus](http://cgit.freedesktop.org/systemd/systemd/tree/NEWS?id=2d1ca11270e66777c90a449096203afebc37ec9c#n57) 270 * As of 2015-07-31 kdbus is not in the mainline Linux kernel. systemd 271 made kdbus non-optional in its release. The kernel maintainers are 272 still debating the kdbus ABI or possible alternatives, but if 273 systemd depends on the current state of kdbus the kernel maintainers 274 are faced with the hard decision to either break Fedora userspace or 275 accept the current kdbus proposal into the kernel with its security 276 and maintainability issues. This is the best example how 277 systemd is forcing you into decisions. Of course if you are a mindless 278 bureaucrat it helps you to keep your job. 279 * [readahead removed](http://cgit.freedesktop.org/systemd/systemd/tree/NEWS?id=2d1ca11270e66777c90a449096203afebc37ec9c#n950) 280 * The first thing swallowed in on Fedora was readahead. Now that (of 281 course!) everyone is using an SSD (at least the developers of 282 systemd do that) it can be removed. Why was it there? Is it possible 283 to make it a separate tool again? There's no time for that, we are implementing 284 new features. 285 286 from 2015-07-31 287 288 Practical systemd 289 ----------------- 290 291 Here is what happens on a stock Arch Linux system, powered by systemd, 292 when a non-root user tries to restart the system: 293 294 $ reboot 295 Failed to set wall message, ignoring: The name org.freedesktop.PolicyKit1 was not provided by any .service files 296 Failed to reboot system via logind: The name org.freedesktop.PolicyKit1 was not provided by any .service files 297 Failed to talk to init daemon. 298 299 In contrast, here is the equivalent error message on a system powered by 300 runit: 301 302 $ reboot 303 init: fatal: unable to create /etc/runit/stopit: access denied 304 305 And on the oldest and best, Slackware: 306 307 $ reboot 308 reboot: must be superuser. 309 310 systemd is driving "just google the problem" attitude, because how the hell are 311 you expected to troubleshoot this kind of error otherwise? 312 313 Further Reading 314 --------------- 315 * [Without systemd](http://without-systemd.org/wiki/index.php/Main_Page) 316 * [Arguments Against systemd](http://without-systemd.org/wiki/index.php/Arguments_against_systemd) 317