Controversial Lennart Poettering Finds His Place In Linux Community - InformationWeek

InformationWeek is part of the Informa Tech Division of Informa PLC

This site is operated by a business or businesses owned by Informa PLC and all copyright resides with them.Informa PLC's registered office is 5 Howick Place, London SW1P 1WG. Registered in England and Wales. Number 8860726.

IoT
IoT
Comments
Controversial Lennart Poettering Finds His Place In Linux Community
Newest First  |  Oldest First  |  Threaded View
JayArby
50%
50%
JayArby,
User Rank: Apprentice
6/28/2016 | 10:31:07 AM
Correction
Mir display server was not an alternative to systemd. There is no relation between the two (you're thinking of the Mir vs. Wayland debacle). Canonical's alternative init system was called upstart.
T.J.D781
100%
0%
T.J.D781,
User Rank: Apprentice
5/11/2015 | 1:39:35 PM
kfv653: Error in article regarding Ubuntu ,Systemd and Mir
With respect, as a Linux user for almost 20 years, and a developer for over 25 years, I must disagree on almost every point you make. 

I read your discussion about Linux very carefully, and I cannot impress this enough: anyone who claims Linux is a desktop OS is "deranged" or living in a dream.  Linux was never intended to support "out of the box" commercial software NOR is it intended as a desktop operating system. 

In order to do either of those things, Linux would have been designed so companies could compile their Windows and other applications on Linux. It's not possible at the present time without extensive changes. Microsoft made certain of that when they withdrew support from the POSIX standard in 2000. The only exception is .NET code, which Microsoft recently opensourced, and that has been depreciated in favor of Windows Runtime as of Windows 8.  
 
If you want a desktop OS with "shrink-wrap" software, use Windows or MacOS.  It is what they are designed for.  You will be a lot happier.  Linux is designed first and foremost to be a UNIX server operating system.

 

As for your points regarding systemd:

 "People who want to have an older style init can still have it that way be configuring it how they want, SystemD supports the old SysV init as well as its new facilities."

 Sorry, but that is only partially true.  Systemd is 'mostly' compatible but will break a number of System 5 scripts. Systemd notes this in their own documentation.  For example, there are no inherited variables.  Systemd is not a "drop in replacement" for System 5.

"The way things were before is that it was impossible to do this."

Not true.  There were several service daemons to do these tasks.  Systemd just replaces all or most of them under a common framework, which consequentially also makes the software incompatible with everything except Linux.  For that reason, I do not consider it to be a good outcome.

"THe people who oppose it claim to be system administrators, they must not be very good system administrators if they find it so difficult to understand the basic model of systemd."

No, actually what they object to is combining all of the server services into a compound system.  The Systemd daemon itself is not as much the problem as the fact that its developers seem hell bent on replacing every service daemon with one that depends on systemd to work.  This adds attack surface to the operating system, e.g. if systemd is vulnerable, then all of the services that use systemd code are also.  It creates a point of failure, because it not only initializes the operating system, but actively manages daemons in session.  If the management portion of systemd crashes, the whole system goes down, and needs a hard restart.   

 
"If they claim to be system administrators, why dont they just use the system V failities to run their jobs, if thats what they want?"

Because they do not work properly in many cases, under systemd. See above or read systemd's own wiki.


 "I think what they are oppoising really is other administrators being able to actually have some control over initialization because previously with the Linux kernel and init it was hard to hook to certain system events and trace what is going on in the system. It was sort of a black box that administrators had little introspection over."

 Actually in my experience, syslogd solved that problem very well.  Systemd's binary logs have had a history of corruption problems.  The developers' response?  Restart the system and erase the logs because they are useless.  Doing that means that you won't find out what happened.  Sorry, but no.  I will use systemd on a personal desktop, but never on a server when keeping a paying customer hangs on whether I can fix a problem.

THis is because systemd DOES NOT take away anyones options it only adds additional options ..."

No, systemd is not the problem.  The Linux distributors are the problem.  They are the ones forcing their users to adopt systemd, rather than leaving the decision to the end user to decide what they need.  THAT is what is making people angry.

The only common versions of Linux that do not force their users to use systemd are Slakware and Gentoo, who have it as an optional install.  I would include Debian in that list provisionally, but future versions may not leave the option to not use it open.

"I think they are doing immense damage to the operating system because of this. I seriously question the intent of these people."

I think they have every right to complain when systemd causes real world problems.  Just as they have every right to complain when Windows crashes.  It does not "damage" anything to say that a problem is a problem.  I'd consider it to be far more damaging that the systemd developers ignore criticism as they do. It took Linus Torvalds suspending Key Sievers' merging privileges for causing kernel crashes before Sievers "got the message".

 
Charlie Babcock
50%
50%
Charlie Babcock,
User Rank: Author
5/11/2015 | 1:01:29 AM
In Ubuntu, systemd displaced Canonical's preferred Upstart.
I have read about two modules developed by Canonical for its Ubuntu distribution that it was forced to back away from. Randy P533 is right. I should have referred to Upstart as the one that systemd displaced.
kfv563
0%
100%
kfv563,
User Rank: Apprentice
5/10/2015 | 6:28:01 PM
Re: Error in article regarding Ubuntu ,Systemd and Mir
I do not believe everyone hates systemd, in fact it seems to be a small, but loud, minority that do. I have worked with systemd and have found little issue with it. People who want to have an older style init can still have it that way be configuring it how they want, SystemD supports the old SysV init as well as its new facilities. systemd is sort of independensable and can be of great utility in being able to schedule jobs to be able to run on certain system events. The way things were before is that it was impossible to do this. systemd has give us only more flexibility and control over the system, and people who oppose it seem to oppose the administrator being able to have the control to be able to configure events in such a fine grained way that systemd allows.

I really don't know what the problem is. I didnt find any problems with the concept, its actually pretty useful. THe people who oppose it claim to be system administrators, they must not be very good system administrators if they find it so difficult to understand the basic model of systemd. If they claim to be system administrators, why dont they just use the system V failities to run their jobs, if thats what they want? I think what they are oppoising really is other administrators being able to actually have some control over initialization because previously with the Linux kernel and init it was hard to hook to certain system events and trace what is going on in the system. It was sort of a black box that administrators had little introspection over. 

Because systemd just provides additional facilities rather than forces anything on anyone, there really is not way to oppose it because people who want things to use old system V type init can do so, so thats obviously NOT what the uproar is about, those who oppose it seems to be trying to stop Linux from being any more powerful or flexible, they are te ones who oppose choice and flexibility in configuration because they want to deny the ability and customizability that systemd provides. THis is because systemd DOES NOT take away anyones options it only adds additional options so by opposing it they are trying to force thei own way of doing things on others. I think they are doing immense damage to the operationg system because of this. I seriously question the intent of these people. 

 

Another fact is while Linux people claim to be so superior and immensely powerful, and the OS to basically to have always been this perfectly useable thing the fact is that even for system administrators Linux has not been a particular flexible operating system for many years in terms of the out of the box software that was available, many Linux systems were down right inflexible and difficult to configure, and the OS has long suffered from a lack of well developed desktop applications, making it almost useless as a desktop OS. It had an HTTP server but this is far from really being able to do a serious server development platform, database systems for instance with failover support necessary in an enterprise environment were a long time coming. MySQL always lagged well behind Oracle. It was good enough for a small web site but when you are talking massive operations where downtime is not an option, it would fall way short. The point of the criticism is that the systems need improvenent but, I am making a point that whenever anyone does try to do something to make the system not so rigid that they are shot down by these people who have this crazy idea that Linux in 1992 was the most wonderful thing and that somehow not having a database with transactions and failover, redundancy and systems necessary for high availability was a wonderful thing. These people basically do not care about anyone else and any other use cases but their own and their own are often quite narrow purposed.
T.J.D781
100%
0%
T.J.D781,
User Rank: Apprentice
5/9/2015 | 3:24:38 AM
Systemd and Mr. Poettering
With respect to your point of view, systemd's reputation with the community has suffered because systemd has caused real problems.

Firstly, it has caused problems with the Linux kernel itself.  This would be similar to someone writing a driver that crashes Windows to the "blue screen of death."  Generally, that is counterproductive, and when asked to fix systemd, its creators attempt to ignore responsibility.  Linus suspended Kay Sievers for this, and in my opinion, it was justly deserved.

The second concern is compatibility. In spite of claims to the contrary, systemd is not entirely compatible with System 5. Nor is systemd compatible with itself. Like so many undeciplined efforts these days, systemd does not guarantee a minimal fixed interface standard regardless of version.  This means that a single version update to systemd can cause your system become inoperable.

Third, Unix code is typically written to the POSIX standard, ensuring that it can easily be ported from one operating system to the next.  Systemd does not follow the POSIX standard. That in itself is not a problem.  There are many Unixes that  use code that violates the standard.  The problem is that projects such as Gnome are now building themselves around systemd, which makes them either crippled or unusable on anything except Linux. That flies directly in the face of the longstanding tradition of the Unix roots of the opensource community, and deliberately tying what was once generic code to Linux only has angered a great many.

Fourth, Systemd's goal is overreaching and forces homogenization.  In spite of what people howl, that it replaces init is of no concern.  Init is not actually part of the POSIX standard.  The problem is that the systemd project has been consistantly attempting to subsume other projects into itself, and suffers from poor project management.  Systemd makes legacy software incompatible, with no concern as to use cases.  It forces every Linux system to use the same software or become incompatible.  This might sound like a good thing to some, but one of the reasons that Linux is more resistant to being tampered with over Windows is diversity.  A diversity that the creators of systemd have set themselves to eliminate.

Lastly, systemd's precieved advantages can be acheived through less invasive means, as other projects, such as uselessd have set out to prove.
RandyP533
50%
50%
RandyP533,
User Rank: Apprentice
5/8/2015 | 8:59:22 PM
Error in article regarding Ubuntu ,Systemd and Mir
Yea.  Good article but confidence in the  authors basic linux knowledge is low.....

There is a major difference between an init system and a display server.  Article states "Shuttleworth already had commissioned his own init alternative in Ubuntu, Mir display server".  The fact is Shuttleworth had commissioned his own init alternative in Ubuntu, --- Upstart --, and lost out to Systemd.  He has in fact commisioned his own Display server called MIR.  The rest of the community is appearing to put their support for Wayland as the next generation Display server, so Shuttleworth may be 'shut out' again.  

 
Charlie Babcock
0%
100%
Charlie Babcock,
User Rank: Author
5/8/2015 | 7:44:01 PM
VMs were hard to get into the kernel; containers too?
The merit of the code rules in the Linux community, right? So why have Poettering and Sievers taken so much heat? In my view, the Linux Kernel maintainers, especially Torvalds, were slow to develop any interest in virtual machines, and it took an artful second try by Qumranet to get a hypervisor (KIVM) folded into the kernel's operations in 2007. That meant the Linux community was five years behind VMware's proprietary ESX Server 1.0. Now containers have come along, with the initiative to get them included in Linux operations coming from outside the Kernel development process, primarily from Google. Understood that systemd, the initialization system that's best for containers, is non-conforming in several ways. But the merit -- the usefulness -- of the code still rules, right? 


State of the Cloud
State of the Cloud
Cloud has drastically changed how IT organizations consume and deploy services in the digital age. This research report will delve into public, private and hybrid cloud adoption trends, with a special focus on infrastructure as a service and its role in the enterprise. Find out the challenges organizations are experiencing, and the technologies and strategies they are using to manage and mitigate those challenges today.
Commentary
Augmented Analytics Drives Next Wave of AI, Machine Learning, BI
Jessica Davis, Senior Editor, Enterprise Apps,  3/19/2020
Slideshows
How Startup Innovation Can Help Enterprises Face COVID-19
Joao-Pierre S. Ruth, Senior Writer,  3/24/2020
Commentary
Enterprise Guide to Robotic Process Automation
Cathleen Gagne, Managing Editor, InformationWeek,  3/23/2020
Register for InformationWeek Newsletters
Video
Current Issue
IT Careers: Tech Drives Constant Change
Advances in information technology and management concepts mean that IT professionals must update their skill sets, even their career goals on an almost yearly basis. In this IT Trend Report, experts share advice on how IT pros can keep up with this every-changing job market. Read it today!
White Papers
Slideshows
Twitter Feed
Sponsored Live Streaming Video
Everything You've Been Told About Mobility Is Wrong
Attend this video symposium with Sean Wisdom, Global Director of Mobility Solutions, and learn about how you can harness powerful new products to mobilize your business potential.
Sponsored Video
Flash Poll