dinit VS minibase

The author has a high-level overview doc here: https://github.com/davmac314/dinit/blob/master/doc/COMPARISO...

Now, of course, with service management the devil is in the details, so before you use something as system-wide init, I find it useful to ask yourself: How do they restart services? Do they give up at some point? How do they notify administrators of failures? Do they detect crashloops? How configurable is the logger? What CLIs are there to debug the state of the system (which service was first to fail, where is its definition)? How to make ephemeral or parametrised services? How to add pre-start, post-start, pre-stop, post-stop hooks? Can you use environment variables in your commands, and where do they come from?

I don't think you can find answers to some of these questions in docs. Once you do learn the answers, they may be disappointing - I indeed found myself quite disappointed in systemd after having to debug many failed-to-boot machines. With s6, I never run it, but there are few choices that raise eyebrows, e.g. the restart delay is hardcoded in source code to be 1 second. With dinit, I have yet to finish reading all its manpages, but at least timeout and restart policies are configurable.

Not to be confused with https://github.com/davmac314/dinit

Still, I applaud efforts like s6 and Dinit as competition is a good thing in general. I hope they'll continue to be improved upon until they've become viable alternatives to systemd for most users.

You can download dinit from github https://github.com/davmac314/dinit. (also read everything about it) Do a simple make && make install which should install it to /sbin/dinit No need to remove systemd or openrc. /sbin/init should be symlinked to whatever init system you use. Read the instructions on dinits page. All the services go into /etc/dinit.d. And you can "dinitctl enable servicename" to enable it. I threw the services I have on my system up at https://gitlab.com/fictitiousexistence-public/dinit-gentoo/ You can copy them and enable / disable whatever you need. Most services I jacked from artix since they have a supported instance of dinit.

It got mass-adopted while being imperfect, so that's to be expected. Thankfully its inception and the criticism that followed have paved the way for the likes of dinit and s6.

this page by Davin McCall, creator of dinit.

I use dinit do manage services on my home server. One of them is Caddy, that shares TLS/SSL cert state with my remote server by using Redis on said remote server. However, since this means that I need to have established a remote connection first before starting Caddy, I would like to know of a method to check if tailscale has in fact finished connecting.

I've used a ton of improvements but I know this community knows a lot more about some parts than me. So here I come. I'm currently using efibootmgr to create an efistub to bypass grub. Using base-minimal with ncurses so terminal apps work. Also I use this in my /etc/dracut.conf.d/local.conf: hostonly="yes" omit_dracutmodules=" network plymouth " And that works perfectly. Don't know exactly what network or plymouth are for but they are not strictly required. Also I use dinit: void_dinit, dinit. This works perfectly fine and works like expected after a little troubleshooting. Also for making dinit about 2x faster is swapping coreutils to sbase. I only use ls from gnu coreutils which is something I may want to switch but I haven't come to finding a good alternative to ls that has colors (exa exists but is 11x as big) Things that I could think of that could improve stuff are: kernel, maybe initramfs. Maybe something else I've looked over. What is all made by GNU anyway? I may have forgot something Void Linux uses.

systemd: Yes; it's awaiting its "PipeWire". Thankfully, the likes of s6 and dinit are very promising. Though I can actually appreciate that systemd is addressed. As ultimately it helps in raising awareness that will benefit whatever software will replace it eventually.

Arguably this is less troublesome to solve compared to the other concerns. As we're inevitably waiting for the system supervision suite that will be to systemd what PipeWire has been to PulseAudio. I'm very optimistic about this as both s6 and Dinit are shaping up lovely.

I remember simple use cases for clone() such as spawning child processes with just enough shared resources to execve(). I remember reading a lot of old emails from Torvalds about it, can't find them anymore.

I used to value portability but now I believe in using Linux everywhere and for everything. I like OpenBSD too but Linux is the stable one you can build anything on. What I wanted to eventually accomplish is a 100% freestanding Linux user space with no libraries at all. Maybe boot straight into the program I want to use, just like we can pass init=/usr/bin/bash in the kernel command line. How far could this go? Using nothing but system calls it's actually possible to get a framebuffer and use software renderering to draw some graphics. I'm guessing pretty far.

By starting from scratch like this it's possible to fix all the historical problems with our systems. For example, I think it's unacceptable when libraries keep global state. This can't be fixed without getting rid of libc and its buffers and caches and errno. Removing this cruft would actually simplify a threads implementation. And then there's completely insane stuff that should be dropped like .init and .fini sections:

https://blogs.oracle.com/solaris/init-and-fini-processing-wh...

A similar statically-linked user space project I found years ago:

https://github.com/arsv/minibase