Sysvinit Future

[chiguy1256]

I respect MX Linux decision to not use systemd. I haven't had any issues with it on other distros that I use. I have not had any issues with sysvinit as well. Is MX Linux going to stick with sysvinit or has there been some considerations for a replacement other than systemd? Just curious. Thanks.

[dolphin_oracle]

I respect MX Linux decision to not use systemd. I haven't had any issues with it on other distros that I use. I have not had any issues with sysvinit as well. Is MX Linux going to stick with sysvinit or has there been some considerations for a replacement other than systemd? Just curious. Thanks.

I don't know how the other devs feel about it, but my intention would be to continue to straddle the sysV/systemd line as long as practical.

[Richard]

Although I had problems with systemd on a couple of other distros before finding MX Linux in April-2014, I continue
to be satisfied with MX Linux as currently conformed only desiring it's continued success and development.

I suppose the question really could be, can MX Linux continue to be "a midweight OS designed to combine an elegant
and efficient desktop with simple configuration, high stability, solid performance and medium-sized footprint"
under the spectre of world domination inherent in systemd that many believe exists. Perhaps that image is not proven
to everyone's mind but I believe it is a part of the reason for the existence and success of MX Linux.

I have read in the forum of some who run MX Linux with systemd with apparently no problem.
So we know it is possible.
Question is whether it is desirable?

[Jerry3904]

I'm with DO on this, especially given the update work now going on.

[chiguy1256]

There are alternatives to systemd and sysvinit. Such as OpenRC, which is used by FreeBSD (TrueOS) and Gentoo.

[anticapitalista]

There are alternatives to systemd and sysvinit. Such as OpenRC, which is used by FreeBSD (TrueOS) and Gentoo.

Have you tried it on MX Linux? Please give feedback if you have or intend to do so.

[BitJam]

There are alternatives to systemd and sysvinit. Such as OpenRC, which is used by FreeBSD (TrueOS) and Gentoo.

I use Gentoo. One of the reasons I chose it instead of Debian, over 15 years ago, is I much preferred their init system over what Debian was using. I'd LOVE to switch us over to OpenRC since it makes so much more sense compared to Debian's SysV init. Actually, I think Debian got rid of some of the worst problems before they abandoned it. But here is the problem with switching to OpenRC: the init.d scripts are different. Right now (thank goodness) Debian packages still come with SysV init scripts. This means we don't have to repack every bloody Debian package that has an init.d script.

I can imagine creating tools to help with this but since we are not Trying to Take Over the World! [TM] I don't think the benefit is worth the effort. Usually, changing the init system goes along with changing the packaging system. That's what happened with Gentoo. If you are creating all new packages anyway then adding new init files is not a big deal.

I'm not saying a clever solution is impossible. I just don't know what it is. If you can figure it out then more power to you! If it is not too onerous then we may want to consider it.

[sanlav]

i will go with D-O opinion with some clarifications :

1. sysvinit was Proved at least few hundreths million times till now. Yes, it is an old serial system but he does his job very well ...

2. a simple search for " systemd problem " gave me 1.5 mill results ... starting with " Linux: Why do people hate systemd? " . Systemd could be good in the future at paralellisation of processes but not now.

3. I could sacrifice few seconds of boot time for stability and assurance that my system work as I though it would.

I rest my case...

[kernelkurtz]

... Is MX Linux going to stick with sysvinit or has there been some considerations for a replacement other than systemd? ... .

... to continue to straddle the sysV/systemd line as long as practical.

I think there's wide agreement here on that. But, of course, someday it won't be practical any more.

I've been happily running MX for most of a year, but this truth has certainly been in the back of my mind that whole time.

So come summer I'll likely drift down the stack to AntiX at least, and keep my eyes on the other evolving options.

My sense is that in the long term I'm going to have to be a lot smarter about Linux internals, and not depend on a turnkey solution with a pretty installer to make it all right. Which in most ways if perfectly fine with me.

[timkb4cq]

My problem with systemd are its failure modes. It took over two years to fix a problem where under certain conditions the manual mount of a partition would be show as succeeding (exit 0) but systemd would quickly & silently unmount it (it did appear in the logs but not in the terminal). The silently is what is unacceptable. It didn't hit a large percentage of users, but was a troubleshooting headache for those it did.
You still have to restart the systemd daemon if you change the UUID of a drive listed in fstab or it won't mount.

Sysvinit doesn't try to do as much as systemd does so it has fewer failure modes and they are more straightforward to recover from.