Linux Distros Without Systemd (2019) (ungleich.ch) 245
New submitter Nico Schottelius writes: It's 2019 -- who has switched to systemd, who hasn't and what can I use if I don't like systemd? Here's the answer in short.From the blog post: If you are reading this post you're very much likely not a fan of systemd already. So we won't preach on why systemd is bad, but today we'll focus more on what are the alternatives out there. Our approach is obviously not for settling for less but for changing things for the better. We have started the world after systemd project some time ago and the search isn't over. So what are the non-systemd distros out there? The author makes a case for why you should consider the suggested distros, but here's the list: Devuan, Alpine Linux, Artixlinux, Void, Slackware, Gentoo, and GNU GUIX.
systemd or not, Linux is garbage. (Score:2, Insightful)
I think I'm about to lose my mind over how everything that resembles desktop computers are slipping further away by the day, with nothing but hardware-compromised junk machines and garbage OSes like Windows 10 and all these Unix-based trash and Haiku (not even worth commenting on) as the only "alternatives"... It's just no fun anymore. It's pure terror. Every day.
Computing is over.
Re: (Score:3)
I was going to just mod you "insightful" but this deserves a reply. I agree with what you wrote, and I think the problem is that programmers are looking to implement the new flashy thing instead of shoring up the problems in the existing software. I've had good experiences with OpenBSD because they focus on stability. Heck, there's no Bluetooth driver in it because it's too complicated/closed-source to implement!
Re: (Score:2)
I've had good experiences with OpenBSD because they focus on stability. Heck, there's no Bluetooth driver in it because it's too complicated/closed-source to implement!
An advantage of Linux is that it's one OS to rule them all. You can run it on a watch, or a router, or a phone, or a console, or a supercomputer, and there are many tools in common from one end to the other. All the BSDs are similar, but you're still going to need a variety of them to cover all your needs.
Re: (Score:3)
That's actually not an advantage. It's absurd to think that one operating system is right for everything. Inevitably, you end up with a compromise that is shit on every device. There is no successful operating system for a mobile device that also has a significant presence in the desktop world, or vice versa. Android and iOS both share their kernel with a desktop OS, but that is about it. Most of the rest of the OS is bespoke for the device.
It's also not true that Linux is one operating system. Leaving asid
Re: (Score:2)
Linux is not an operating system. The Linux kernel does the things needed by all of these differing applications, and an appropriate userland gets glued on as needed. The kernel can be customized, leaving out parts that are undesirable in certain situations, but convenient similarities remain. One does not have to learn the peculiarities of the kernel time and again, as one does when switching BSDs.
Of course it's an advantage.
A late friend and software developer (creator of omniremote, among other things) o
Re: (Score:2)
(Also, yes, I realize I conflated kernel and OS in my original comment. Sorry.)
you are deeply flawed. (Score:2)
No, you are deeply flawed.
Software does not work and is not developed like Hardware.
Re: (Score:2)
Unfortunately, this "rot" has now reached Linux as well. And the biggest problem is that all these young, inexperienced and often very arrogant coders usually do not do good maintenance and sometimes do not even finish their product. My guess is that overt time all that crap (including systemd) will get removed again, and that the current hype of containers and VMs will mostly go away. It does make things easier to deploy, but much harder to administrate and that is a net loss.
Re: (Score:2)
The hype of containers and VMs will go away, but containers and VMs won't go away. Just too darn useful.
Re: (Score:2)
They are useful, but they are only useful for special situations.
Re: (Score:2)
Such special situations happen to consume the majority of my working day, can't speak for others.
Re: (Score:2)
and that the current hype of containers and VMs will mostly go away
What sequence of events do you foresee leading to that?
Re: (Score:2)
The Gartner Hype Cycle doesn't predict obsolescence. Its end state is the "plateau of productivity," where technologies function as they're intended and people just get back to work.
Re: (Score:2)
It does make things easier to deploy, but much harder to administrate and that is a net loss.
In what way? It sounds like you're imagining taking monolithic, scale-up legacy applications, shoving them into containers, and then being isolated from their inner workings as an administrator. That's almost universally the wrong way to do things, and if you build container-native applications properly (stateless microservices, scale out, fail fast, etc) the level of automation you can achieve should make life much easier. Refactoring and recomposing existing applications to work this way can be challengin
Re: (Score:2)
Computing is over.
Let's be honest, it was over when the microcomputers arrived.
#Batch4Life!
Re: systemd or not, Linux is garbage. (Score:2, Insightful)
I think he's on your side, dumbass.
Did you not even try wiki? (Score:5, Informative)
https://en.wikipedia.org/wiki/... [wikipedia.org]
Re:Did you not even try wiki? (Score:4, Informative)
The information on the page is more up-to-date than that. For example, the e operating system [wikipedia.org] was added to the category less than six months ago.
The change to the category page on 8 October 2017 was to remove "This category is being considered for deletion." from the page: https://en.wikipedia.org/w/ind... [wikipedia.org]
Re: (Score:2)
Congratulations for demonstrating you don't understand Wikipedia.
FreeBSD (Score:5, Interesting)
In my workplace, we've replaced almost all Linux instances with FreeBSD. It has been a godsend, honestly. There is no longer worry about packages being out of date and relying on PPAs anymore. Kernel + userland + ports all being managed as a single collective set by the same people, rather than each being separate entities all conflicting with one another has enabled a level of application stability we've never seen before in this organization. For all of our major packages, we're usually only a day or two behind upstream provides (the application developers themselves), rather than months/years behind like many Linux distros are with their packages.
Re:FreeBSD (Score:5, Interesting)
So you're the 0.003% that went from Linux to FreeBSD, while the remaining 99.997% went from FreeBSD to Linux. Got it.
A little secret, when you operate within a stable distro then you get central coordination and package development/management that's well internally managed. The problem, which is the problem that led to Ubuntu and the PPA is that people want stuff not available in the stable distro. They start wanting features and before long they're adding backports, migrating to testing, have turned their systems into PPA hell, or have manually downloaded packages and attempted to manually meet those dependencies.
I started with Slackware back in the nineties. It was great, until time had passed and there was no good package management system. Tried SuSE, quickly got tired of additional dependency problems and fairly soon ended up on Debian for many years. Unfortunately now they've trusted someone that couldn't even get an audio subsystem to universally work properly to rewrite init.
I like the idea of systemd. I don't care for its implementation. Having played as an end-user with some projects that were written with it in mind I can see how it has advantages for administration, but even then, I see the same unfinished approach; things in these projects don't internally work either, and project maintainers get hostile when asked about broken features even if the project is five full revs into existence.
And if it's any consolation, having seen some presentations from the Devuan team, in some sense they're right but for the wrong reasons. As bad as systemd is, they fail to acknowledge how modern computing has changed and how paying for virtual machines on someone else's servers with things like spinup time and the ability to quickly personalize VMs booted from generic templates actually do matter.
The geeky musical group Three Dead Trolls in a Baggie were right, every OS sucks.
Re:FreeBSD (Score:4, Insightful)
There is quite a bit more FreeBSD than you may realize. Save for Android, generally Linux distros advertise themselves as much front and center that they use the Linux kernel. FreeBSD distros on the hand, often times don't. There are many network appliances that now run various forks of FreeBSD. Major hardware vendors such as Dell are on-board with FreeBSD. Major content providers such as Netflix are on board with FreeBSD. macOS borrows quite a bit from FreeBSD too. Plus there is the entire homelab community rallying behind FreeNAS, pfSense, and OPNsense. There is just significantly less chatter and online posts from people that use FreeBSD, because for the most part, at least in the server and network appliance space, it JUST WORKS.
Re: (Score:3)
FreeBSD is a miniscule fraction of all servers and embedded, compared to Linux. In absolute numbers, certainly enough to keep the project going but in no danger of achieving global domination, unlike Linux.
Re: (Score:2)
On the 'scale' FreeBSD runs Netflix [phoronix.com][0] and WhatsApp [wired.com][1]. It also powered Hotmail. [wikipedia.org]
For the small scale "embedded" (depending on your definition of embedded) it powers the PS3 [psdevwiki.com][2] (>87.4 million units) and PS4 [wikipedia.org] (96.8 million )
Stop looking at 'percentages'. This day in age almost any popular project is at critical mass for being self-sustaining. FreeBSD has more than enough critical mass to continue. Even if its total percentage shrinks (as other OS get introduced and IoT). I bet Linux looks like a small fracti
Re: (Score:2, Interesting)
Facebook is a Linux shop, so it won't be long before FreeBSD is swapped out for Linux, it just makes no sense to do otherwise. Didn't I already remark that FreeBSD is self sustaining? Why yes I did. However, FreeBSD will never dominate, and that is a good thing, because FreeBSD provides no protection against being taken private, which would offend the public interest.
Look, if this bothers you then Fork a GPL version of FreeBSD, then you can compete on the merits. And lose anyway, but that's another story.
Re: (Score:2)
To counter your possibly biased subjective opinion, try a swig of objective reality. [solvedns.com]
Sorry, reality is a thing, rhetoric won't change it. It's hardly a bad thing that Linux achieved world domination instead of FreeBSD, the latter being so easy to take private.
Re: (Score:2)
Linux hasn't achieved World Domination.
You can't take FreeBSD private, it's Open Source.
Re: (Score:2)
Re: (Score:2)
I like the idea of systemd. I don't care for its implementation. Having played as an end-user with some projects that were written with it in mind I can see how it has advantages for administration, but even then, I see the same unfinished approach
Guix is worth a look [gnu.org].
Re: (Score:2)
> I like the idea of systemd.
In civilized states, we make such perverts register with the police! :)
hawk
Re: (Score:2)
On FreeBSD I don't have to worry about my ethernet interfaces being enumerated differently after a reboot. Oh wait I need to write my own udev rule to fix that. Why in the fuck isn't it fixed in the first place?
Because the issues are not nearly as trivial as you imagine. If you think that FreeBSD "fixed it" then maybe you don't get out much.
Re:FreeBSD (Score:4, Informative)
Well, driver probing is generally not predictable for modern technology [freedesktop.org] I suspect that your denial wrt other operating systems is based on those systems only have a single, on-board network device, in which case Linux will behave the same.
For what it's worth, for anything short of a server, I find the "predictable network interface names" to be stupid and annoying, and I simply disable them with a kernel parameter and they go back to being good old EthX. BTW, the same EthX on every boot, I'm not sure what you are going on about.
Arguably, "predictable network interface names" should be an install option. I would not go so far as to say it should be off by default. The vast majority of users these days will never see a network device name anyway, it's only nerds like us who obsess over exactly what they should look like.
Re: (Score:2)
The fact that AppArmor has a "complain" mode" should tell you something.
Like what? You seriously can't think of any use cases for complain mode? It's not a toggle for the entire system, you know (unless you want it to be).
Re: (Score:2)
The fact that its called complain mode. Like your application is a nuisance. Oh you you're special and need a exemption.
Re: (Score:2)
On FreeBSD I don't have to worry about my ethernet interfaces being enumerated differently after a reboot. Oh wait I need to write my own udev rule to fix that.
No, you don't. If you want the 'traditional' behaviour (which is broken for many use cases, e.g. if you have multiple NICs with different drivers which can change order, if you have NICs with SR-IOV support, dynamically adding/removing interfaces from VMs etc.), boot with net.ifnames=0, or use one of the other methods to disable biosdevname [freedesktop.org]
Oh, and many people tout FreeBSD as being used by network appliance vendors, but many of them effectively use a similar approach to interface naming (e.g. ).
Re: (Score:2)
On FreeBSD I don't have to worry about my ethernet interfaces being enumerated differently after a reboot.
Systems with sequential boot order tend to be more consistent in terms of hardware enumeration, but its nothing "trivial to fix".
Re: (Score:2)
In my workplace, we've replaced almost all Linux instances with FreeBSD. It has been a godsend, honestly.
You didn't have to switch to FreeBSD to get this, you could have had it on any number of rolling distros, or "unstable" streams of distros:
* Arch
* OpenSUSE Tumbleweed
* Fedora Rawhide
etc.
On on something like RHEL, you could also have considered Application Streams (RHEL8) [redhat.com] or Software Collections (RHEL7) for a stable OS with up-to-date application dependencies.
There is no longer worry about packages being out of date and relying on PPAs anymore. Kernel + userland + ports all being managed as a single collective set by the same people, rather than each being separate entities all conflicting with one another has enabled a level of application stability we've never seen before in this organization. For all of our major packages, we're usually only a day or two behind upstream provides (the application developers themselves), rather than months/years behind like many Linux distros are with their packages.
Many distros? You mean the two or three LTS-type/enterprise distros out of the > 100 distros? Most distros 6-12 months behind on > 20 000 package
Re: (Score:3, Funny)
My workplace is changed from CentOS to Windows, and it...hold on, I need to reboo......
Do users care? (Score:3, Interesting)
Re: (Score:3, Insightful)
fight isn't over when professional admins have to "do battle with systemd" to get stuff to work, and have a nightmare of troubleshooting on their hands when systemd fails during boot.
this garbage needs to be routed out of the GNU/Linux world, shame on Linus for pandering to this poorly engineered garbage, he's turning into a pussy in his old age.
Re: (Score:2)
Re: (Score:3)
It would be nice if distros like Ubuntu and RedHat would give users a choice upon installation. There's a manual way to switch but I'm a bit leery trying that and breaking an already working system.
Then:
1)You couldn't take advantage of the actual systemd features that were the reason for switching, because the legacy init would't be supportable.
2)You would need 4x the dev/testing effort (maintain both, plus test switching in both directions).
It would be much more efficient to fix any remaining bugs in systemd. Feel free to post any you have filed or contributed to fixing ...
Re: (Score:2)
"The problem is the massive removal of code from software such as KDE (think about suspend/resume management for instance), in favor of systemd, removal of code that was functional and tested." sounds good for KDE and all other DEs - a unified system.
"Sure, it is more efficient to have a monolithic development with a single entry point. Unless it fails, and then everything fails." so do you
Re: (Score:2)
Linus has nothing to do with systemd. The guy everyone hates on got the folks at gnome to rely on it for a few things, and tada... it is now everywhere.
Personally my only complaint is that for what I saw as an end user on a desktop system and a couple of VPS-es I never saw an issue with the old init systems so I'm not sure they really needed replacing, and now it is just another body of knowledge to ignore/forget and a new one to learn. I guess if you want a technical complaint then the binary logging doe
Re: (Score:2)
Re:Do users care? (Score:5, Insightful)
Having spent years hacking init scripts to make third party custom applications run on various Linux distros, I for one welcome systemd. I'd far rather mess with a simple ini file that works well, provides more logging than i had before, and offers process supervision event to the dumbest of daemons. More than once I had to edit an init script and add "set -x" to the top of it to try to trace out just where things were going wrong. Often init scripts were extremely crude, reinvented the wheel, and not quite compatible with the distro I was running on. I very much am glad to have systemd. I also appreciated launchd on MacOS, and even got to sort of appreciate Solaris' SMF system. If old-style init scripts work for you, great. You can still use them with systemd even. But I'm glad to have it simplified.
I've never dealt with any systems involving containers or clusters with dynamic storage devices and hot-swap hardware, but I can certainly see how systemd is really designed for such setups.
I used to hate systemd when I first heard about it. But now I quite like it. I don't interact with it very much, haven't had any problems with it (I still turn on old-school syslog for convenience). It's invisible as far as I'm concerned. I dislike typing systemctl--just hard to type for whatever reason; doesn't roll off my fingers--so I use the old "service" command which still works.
Re: (Score:2)
Isn't it really annoying that everything that systemd replaces ends up needing 3 - 7 more commands to do what the old ones did just fine? It's just so disgustingly similar to Windows' Powershell.
systemctl blah blah
service blah
Uh, normally you would have had:
service sshd restart
Now you have:
systemctl restart sshd
Are you complaining about two more characters?
Also, how did you get the equivalent of:
systemctl is-enabled service
systemctl kill service
On most of the systems I have access to, service foo status results in systemctl status foo anyway. Also, you should install bash-completion ...
ip address show
ifconfig
What has this got to do with systemd? Install net-tools if you want ifconfig, or instead learn the shorter commands for iproute2 (e.g. 'ip addr'
Re: (Score:2)
Re: (Score:2, Insightful)
fight isn't over when professional admins have to "do battle with systemd" to get stuff to work, and have a nightmare of troubleshooting on their hands when systemd fails during boot.
I haven't recently (in the last 3 years) see systemd failing during boot. I have seen systems running systemd fail to boot, but they would have failed the same way with a different init system.
Could you point to any systemd bugs filed upstream about these failures?
Also, I don't know any professional admins who "do battle" with software they need to support, they learn first learn how it works, and if they see areas that need improvement, they do something about it (e.g. contribute fixes, or file bugs etc.).
Re: (Score:2)
Re: (Score:2)
Re: (Score:2)
People who do battle with systemd could just RTFM and move on with their lives. The vast majority of anti systemd army fights out of ignorance not because they have something worth fighting about.
Now you want a real battle, write an init script then fire up a daemon from you init daemon to manage your services because your PID 1 is too dumb to do it's job.
SystemD and Kernel (Score:2)
Systemd has almost nothing (absolutely nothing?) to do with the kernel,
Uses extensively Cgroups / Namespaces / etc. ?
Looking forward to be a client of the in-kernel bus (BUS1) as soon as that thing becomes a feature ? (probably in some of the future 5.x versions)
A huge chunk of the hate that systemd gets it's because it's Linux-specific (as opposed to a generic Unix solution) because it extensively relies on specific features of the Linux kernel (and thus counldn't work on any BSDs. Those are progressively shifting to LaunchD as their own solution to phasing out sysvinit's ya
Re: (Score:3, Insightful)
There is no fight, it is about choice. Some distros removed that choice, and some people DO care.
Do users care what desktop they use? Some do.
Do users care what kernel they run? Some do.
Do users care what distro they use? Some do.
Then, why wouldn't users care what init they use? Its just like every other component in your distro. Some people want to replace X11 with Wayland, that is fine. Some people don't need Pulseaudio because jackd is better, and some people don't need no stinking systemd.
This is not ev
Re: (Score:3)
Most average users of linux want a system that...
You think you can speak for "most average users"? Because you certainly don't speak for me. I'm a professional Linux developer of 20+ years experience (starting with Red Hat Linux 5.0) and things like systemd are a big deal for me! I'm a programmer and systems admin/engineer.
That said, although I have been somewhat slow to adopt systemd, I'm starting to get used to it and I'm actually starting to like it. Things that used to require hackish software like xinetd are really easy to rewrite under systemd. Bare
Re:Do users care? (Score:4, Interesting)
At least two issues are philosophical. Part of the UNIX philosophy is that every program should do exactly one thing and do it well. Systemd definitely does more than one thing and arguably doesn't do them well.
Systemd writes log files in binary format instead of ASCII, which goes against another part of the UNIX philosophy to prefer plain text interfaces because they are universal.
Another major issue is choice. Many people would have preferred not to switch to systemd, but nearly all Linux distributions switched to systemd and refuse to support alternatives.
Also, the systemd developers have a track record of marking significant bugs as WONTFIX. If everybody thought systemd was the greatest thing ever, that still would have been problematic. Since many people didn't like systemd to begin with, it was often the last straw.
Re: (Score:2)
Except that he got an excellent and accurate response before you even tried to suggest he wouldn't.
PCLinuxOS (Score:2)
I'm surprised this one wasn't mentioned. It seems to be a somewhat popular option. It's the one I settled on, when I set up my laptop for dual-boot a couple of years ago.
https://www.pclinuxos.com/ [pclinuxos.com]
Re: (Score:2)
PCLinuxOS was created by Texstar. Before PCLinuxOS, Texstar ran a very well respected repository for Mandrake/Mandriva.
OpenWRT/ / LEDE (Score:3)
OpenWRT / LEDE does not use systemd.
But maybe it does not count as a distro.
Although I think it should, because it is quite useful for some purposes.
Gentoo (Score:3, Informative)
I used Gentoo a long, long time ago. Then I switched to Sabayon because it was like Gentoo but the software packages are pre-compiled so I didn't have to spend so much time and trouble having them compiled on my own machine. But once Sabayon like most other distros switched to systemd, I tried it and gave up on it. For the past few years I have been back on Gentoo. Gentoo is fast and gets compiled specific to my hardware. Gentoo gives users the choice of systemd or not (I choose not to use systemd). Linux distros *should* be about choice, dammit! Plus, I am now using ZFS as my root filesystem, which is great. Compiling the packages locally is still somewhat of a pain, but for the most part I am very happy with Gentoo. I won't use systemd no matter how many bugs they fix. To me, systemd is too much like something big and bloated you would find in Windows.
Re: (Score:2)
Great post, thank you. +1 informative.
To me, systemd is not a set of scripts that I can edit if needed.
To me, systemd is overly automated- tries its best to do what it thinks is needed. Wizards. Too difficult to control, prevent it's troubles.
And frankly, I just don't need it!
Re: (Score:2)
I'm in full agreement with you.
I've been using Gentoo since 2003, I did try systemd once and found it to offer me no real advantages over openrc. I have openrc working in parallel startup modes, all the machines I've built in around the past 5 years had SSDs for their root filesystems and I therefore never saw any noticeable speed differences between systemd and openrc.
What administration I have done in systemd (my day job is Linux server security on mostly Red Hat systems with systemd) seems both illogical
Slackware (Score:3)
The oldest, still the best.
Go Slack!
(Seriously, though, it's worth using -- just plain Linux, no frills, does the job)
Re: (Score:2)
I wanted to get away from Fedora's systemD so I went to Slackware for a new system. The package system is archaic and many programs I like aren't available. The additional programs available from slackbuild don't always build and I was unable to make some build. even after tediously chasing down dependencies. Dependency hell was alleged to be closed a decade ago, but it's still with us.
Next I tried Devuan. I can't remember what went wrong there, but after two weeks struggling I gave up. Now I'm back with F
Re: (Score:3)
Which Slackware version did you try?
Which packages do you need that wouldn't build?
Re: (Score:2)
I run Slackware since ... before 4.x (?) ... and I'm not going to switch. However the 15 is overdue IMHO. KDE/VirtualBox/digiKam all moved to Qt5 long time ago ... thanks God for Eric and his ktown repository.
Also I had big trouble compiling pidgin with SameTime support - i.e. mostly the Meanwhile library [sourceforge.net] that implements Sametime protocol. Actually it still does not work here. Akonadi is unstable. Telepathy cannot connect to Google/Skype/ICQ, KDevelop semantic analysis is buggy, my favorite KDE theme eats a
A stop job is running (Score:2)
I see a lot of "a stop job is running" from systemd, with no obvious way to diagnose or bypass. I would call that a severe flaw, but fixable. Odd that there aren't more complaints about this, and pressure on devs to address it. Otherwise, systemd has not been the end of the world for me, useful even. I worry more about the project governance than the quality of the software itself.
Re: (Score:2)
It's a design defect. Three step solution: 1) Think deeply 2) Write, debug and submit patch 3) Lennart gives you the finger, rewrites it, merges his only trivially different version. Congratulations for a job well done.
Seriously? (Score:3, Interesting)
And this is why Linux is still a pile of crap in 2019. This systemd thing is more idiotic than the republics who still can't let go of Hilary's emails.
You know what list I want? A list of distros that don't have a melt down when you plugin in an external monitor.
I want a list of distros that can dynamically switch between integrated and discrete graphics on a per-app basis. You know, like Windows and MacOS has been able to do for over a decade.
Or maybe a list of distros that properly support basic functionality like keyboard hotkeys.
Or maybe a list of distros with a GUI that is actually intuitive to use and facilitates efficient workflow (ie: not gnome), and is stable enough to not have landmines that can take down the whole desktop (ie: not KDE). While I'm asking for the moon, it would also be nice if scaling was properly passed to applications automatically so I didn't have to manually edit the scaling on almost every application I use so it doesn't render the size of the a postage stamp.
But who am I kidding. The linux community cares more about systemd, colour schemes, and how 'pretty' linux looks. Having an actual robust OS is secondary. And then they wonder why nobody uses linux.
Re: (Score:2)
Re: (Score:2)
Re: (Score:2)
You know what list I want? A list of distros that don't have a melt down when you plugin in an external monitor.
Poettering will work on that next. And just like when distros stopped melting down when you plug in headphones Linux users around the world will cry foul. If you can't do the etc dance while hacking the gibson to get your external monitor to work at the correct resolution you're just not Linuxing right.
Big fork or sub-forks? (Score:2)
Since some swear by systemd and others swear at it, it probably will not go away. Each may fit different environments and/or personalities.
Will Linux fork into "Linux" and "Linuxd", or do you believe each distro will start offering a systemd version and a non-systemd version?
Serious question (Score:2)
I've not closely followed the systemd discussions, but as a basic, everyday-at-home Linux user, what are the implications of system for people like me?
I installed Linux Mint, tweaked a few things and then went about my business. I'm not a developer and I don't normally dig around in the internals of my system, so what are the ups and downs or pros and cons for a basic Linux user like me?
Or put another way, why does it or should it matter to a user like me?
Re: (Score:2, Troll)
Now with an actual kernel inside (Score:3)
It doesn't even include the Linux kernel.
It *did* not. :-D )
Now starting with WSL2, they actually included a Linux kernel running in Hyper-V, because Microsoft apaprently sucks at trying to replicate it successfully.
(They didn't even manage to reach the first "E" of "EEE" before given up
Re: (Score:2)
I'm not sure it's so much sucking as it not being a great way to do things.
It's not like Wine is perfect
Re: (Score:3)
Re: (Score:2)
You mean it will someday include GNU/Linux running in a VM
sounds like a solution for losers.
Re: (Score:2)
Knowing that my own Linux server runs several containers, already much Linux is VMs all the way down.
Sounds like a solution.
Re: (Score:2)
Knowing that my own Linux server runs several containers, already much Linux is VMs all the way down.
Running Linux on Windows is like putting Brie on Wonder Bread.
Re: GNU/Windows by Microsoft the best one (Score:3)
How much do I care what the hypervisor is?
Yeah, that much. None are perfect.
Re: (Score:2)
None are perfect.
Linux is less likely to asplode. Who cares about perfect?
Re: (Score:2)
containers aren't virtualization, irrelevant
sounds like you're unqualified to know a loser when you see one
Re: GNU/Windows by Microsoft the best one (Score:2)
Sure, it's a different concept. Yet, indistinguishable to users... Not the admin, the users.
And yet Hyper-V does just fine running a Linux distro within Windows 10.
And no, I'm not able to recall an abend or blue screen on my personal Surface Pro 3. My work PC had two this year, but it's hugely secured and much of that makes it frantic and unstable. And the SP3 gets the fast ring updated, I'm in the Insider fast loop. I've even escaped the known bad patches somehow, probably due to hardware...
Windows 10 get
Logs (Score:5, Informative)
Also; the fubared way it steals logs. (journald does that).
NOTHING, litterally nothing prevents you from running rsyslog alongside it.
Some distro even come preconfigured this way out of the box (e.g.: openSuse Tumbleweed. You get both journald and rsyslog)
Re: (Score:2)
I am a habitual rsyslog user anyway; but still for the life of me can't get rsyslog to work with journald. At least not with tomcat standard out logs.
. . . . yes, it would be nice if my devs would stop putting critical (and PHI) in std.out. . .
Re: (Score:2)
Alpine initial install is busybox, but you can install any of the full packages and the /bin /sbin, etc., links go away. I tend to replace many of them because busybox doesn't support all original command-line parameters. Just: "apk add bash".
Soon I will be running some live servers on Alpine. 5 years ago a friend had a small NAS- I forget the story but it needed to be wiped and reinstalled. I don't remember the brand or distro (or custom? don't remember). I had just started trying Alpine, so I decided
Re: (Score:3)
it is the polite way of saying "massively overpriced".
Re: (Score:3, Interesting)
I like some of the ideas of systemd, but the overall attitude of how it's designed and how Poettering acts, is horrible. He has so little clue that he does not even realize he has no clue.
Yea, Dunning-Kruger far-left poster case. And a lot of people with apparently not much real-world insight that fall for his act.
Re: (Score:2)
Like it just being a bunch of scripts, that you can adapt at will, like a *normal* freaking system glue
You are conflating the systemd infrastructure with the service configuration. Look in /etc/init.d, what do you see? Right, lots of scripts, just like before. And there are other services, that you can adapt at will. You can argue that it is "not normal". True, it is not SysV init, but what is so great about SysV init?
BTW, systemd configure resides in normal files. Not sure what you're going on about. No doubt there are legitimate things to hate about systemd, but so far you are shooting wide of the mark.
Re: (Score:2)
but the overall attitude of how it's designed and how Poettering acts, is horrible.
If the code is good, I don't care about the attitude of the developer.
Re: (Score:3)
I care about developers that mark actual bugs as wontfix for no rational reason.
Re: Third option: systemd from the sane parallel u (Score:2)
Re: Third option: systemd from the sane parallel u (Score:4, Informative)
Re: (Score:2)
in that case you fucked, systemd code is garbage
As if you've ever looked at it.
Re: (Score:2)
Re: (Score:2)
Really. Gentoo documentation historically had a lot of search hits for a substantial cross section of technical issues, independent of distro. You out yourself as a bit of an idiot.
Re: (Score:2)
It sounds to me like you're the one who woke up this morning and discovered only one of his two testicles had dropped into his scrotum.