Good news on the init front

Post Reply
Message
Author
User avatar
Old Giza
Posts: 426
Joined: Wed Apr 16, 2014 10:31 pm

Good news on the init front

#1 Post by Old Giza »

Jesse Smith of Distrowatch has taken it on himself to consolidate patches and improvements to init and get a new release out. Plans are to continue the process and help make init viable into the future.

User avatar
BitJam
Developer
Posts: 2283
Joined: Sat Aug 22, 2009 11:36 pm

Re: Good news on the init front

#2 Post by BitJam »

That looks great! We may have a repo that contains all of our fixes to make SysV init work on Debian. We've been trying to keep all our fixes in one spot so we can share them with others. Is there a way to get in touch with them?

Yes, here is the repo: modified-init.d-scripts. I'm not certain this is 100% up to date but we could easily make it so. Dolphin and anticapitalista may know best. The modified scripts for the installed system are in a separate directory from those for the live system.

BTW: you should use "SysV init" in your title. Otherwise it is very confusing because there are many other init systems like systemd, OpenRC, and UpStart.
"The first principle is that you must not fool yourself -- and you are the easiest person to fool."

-- Richard Feynman

User avatar
asqwerth
Developer
Posts: 7210
Joined: Sun May 27, 2007 5:37 am

Re: Good news on the init front

#3 Post by asqwerth »

There's a user's review of MX 17.1 on distrowatch that claims:

"Run a group of networked Linux desktops. I started looking at new versions of Linux after noticing that the meltdown and spectre fixes were slowing down the systemd computers more than the systems without systemd. ...".

Could this be true?
Desktop: Intel i5-4460, 16GB RAM, Intel integrated graphics
Clevo N130WU-based Ultrabook: Intel i7-8550U (Kaby Lake R), 16GB RAM, Intel integrated graphics (UEFI)
ASUS X42D laptop: AMD Phenom II, 6GB RAM, Mobility Radeon HD 5400

User avatar
Adrian
Developer
Posts: 8248
Joined: Wed Jul 12, 2006 1:42 am

Re: Good news on the init front

#4 Post by Adrian »

I don't know much about the subject, but would guess the most influence if any would be the boot time, after the boot is complete launching and running programs would be pretty much the same regardless of the init system.

User avatar
Jerry3904
Administrator
Posts: 21881
Joined: Wed Jul 19, 2006 6:13 am

Re: Good news on the init front

#5 Post by Jerry3904 »

Jesse talks about this on his Twitter account quite a bit.
Is there a way to get in touch with them?
Which them? I have an email for Jesse that I will PM to you, if that's what you mean.
Production: 5.10, MX-23 Xfce, AMD FX-4130 Quad-Core, GeForce GT 630/PCIe/SSE2, 16 GB, SSD 120 GB, Data 1TB
Personal: Lenovo X1 Carbon with MX-23 Fluxbox and Windows 10
Other: Raspberry Pi 5 with MX-23 Xfce Raspberry Pi Respin

User avatar
Paul..
Posts: 1777
Joined: Sun Mar 18, 2007 6:34 pm

Re: Good news on the init front

#6 Post by Paul.. »

...could be the beginning of something big...great news.

Asus Prime X570-Pro | AMD Ryzen 7 3700X
16 Gig DDR4 3600 | Radeon RX 5600 XT Graphics
Samsung 860 500GB SSDs (2)

User avatar
asqwerth
Developer
Posts: 7210
Joined: Sun May 27, 2007 5:37 am

Re: Good news on the init front

#7 Post by asqwerth »

Jesse has generally taken an even-handed approach to the different init managers on DW. He keeps an objective tone in all his reviews and commentary.

For him to now come out and say he's helping to collate all the new patches and to maintain sysvinit is a significant development.
Desktop: Intel i5-4460, 16GB RAM, Intel integrated graphics
Clevo N130WU-based Ultrabook: Intel i7-8550U (Kaby Lake R), 16GB RAM, Intel integrated graphics (UEFI)
ASUS X42D laptop: AMD Phenom II, 6GB RAM, Mobility Radeon HD 5400

User avatar
Fornhamfred
Posts: 249
Joined: Sun May 02, 2010 12:49 pm

Re: Good news on the init front

#8 Post by Fornhamfred »

Found this on the PclinuxOS forum, I thought it may be the same person.
PCLinuxOS-Forums
Fun => Software Discussion => Topic started by: NoIBnds on March 26, 2018, 07:58:04 AM
Title: SysV init - Init's alive!
Post by: NoIBnds on March 26, 2018, 07:58:04 AM
Saw this, thought it might be helpful. http://distrowatch.org/weekly.php?issue ... 26#opinion (http://distrowatch.org/weekly.php?issue ... 26#opinion)

Now, whether you like SysV init (hereafter referred to as simply "init" to separate it from OpenRC or systemd) you can probably appreciate that I was curious as to who, if anyone, was maintaining init's ageing codebase. There had not been a new, official release of init for several years. Was this because init was small and old with a codebase that got so much attention that any bugs were squashed years ago? Were the developers quietly pushing out patches for old versions without publishing a new release? Was the code simply being left to rot? I decided that I should find out since init was at the core of several distributions I was using.

At first glance, the upstream project appeared to be dead. Development had continued in the code repositories for a time after the last release (version 2.88), but the bug tracker and mailing list painted a picture of a project which had gradually run out of stream a few years ago. However, people had continued checking the code and writing patches. I found that several Linux distributions maintained multiple patches to fix small bugs, clean up documentation or improve performance. People had gone looking through the code and fixed problems. Then, without activity upstream, the developers had added their patches to their distributions' init packages and moved on.

The result was that there were several distributions running init, but some were running slightly different versions of init. Logs from init might look a little different across Linux distributions, manual pages might have different explanations for certain features. While it was good people had been fixing init as needed, those fixes were not being picked up and shared by the rest of the Linux ecosystem.

I wanted to improved on this situation. Relatively few Linux distributions may be running SysV init these days, but projects like antiX, Devuan and others continue to ship the classic init software. Other projects still use init, if only as a jumping off point for OpenRC to run. I thought it would be best if these conservative systems could all benefit from bug fixes, documentation improvements and new features. Since I hoped to unify the various branches of changes that were forming, I wanted to avoid forking init, so I approached the upstream team. They were welcoming and I became the newest member of their team. Then, with some guidance and kind words from developers who had worked on init in the past, I got to work combing through the patches and bug reports.

It was slow going working on an unfamiliar codebase and trying to stitch fixes together. Sometimes distributions had fixed the same problem, but had done it differently and I tried to put together the best solutions from the bug reports and patches available. At the end of February I was able to upload a new beta package of init for testing. Trying out the new beta on various test systems, both x86 and ARM, has gone well and a stable release should be available soon.

There are a few reasons I'm bringing the new work going into init to light. The first is I am hoping Linux distributions will test and adopt the new release that is about to be published. It would be nice to get everyone using the same code with the same fixes. That will hopefully make it easier to distributions to share improvements in the future and for users to submit bug reports.

The second reason is working on projects like this takes time. Working on patches and trying to find (and test) speed improvements on init takes resources and effort and it would go more smoothly if the project was funded, by either individuals, companies or downstream distributions that use the code. The work will get done with or without help, but I hope it can be done quicker with the support of people who are still using init for one reason or another.

The third reason is I would like to invite people to submit issue reports and feature requests to the upstream bug tracker. SysV init was sleeping, but it's not dead and can be improved to suit its users' needs. I think something which put people off using systemd was that it introduced a revolution in init software while many people prefer a gradual evolution of technology, with small improvements rather than radical new designs. I would like to find out what people who are using init, or are considering using it, want from their init software.

EDIT: Sorry did not realise that this post is a copy of the distrowatch item from Jesse Smith.

User avatar
Adrian
Developer
Posts: 8248
Joined: Wed Jul 12, 2006 1:42 am

Re: Good news on the init front

#9 Post by Adrian »

BTW I think that's one the danger for open source projects when people lose interest they go after the shiny new thing they drop maintenance to stuff that actually works. It's great that Jesse Smith is doing this needed maintenance, maybe we can get in contact with him to see how we can package and test his patches, leveraging MX Test repo maybe.

Post Reply

Return to “General”