Today reading my RSS I found out an interesting news that the Blog The Look of the Replicator, and it is that in the Manjaro Community several ISOs have been launched with the peculiarity that they do not use Systemd like init, else OpenRC, the startup system used by Gentoo.
I don't know about you, but the Systemd theme is already touching my balls a lot, and the more I read, the more I realize that although for the end user (or for many) it does not represent anything super relevant, at least to me , I don't like the path this is taking. I believe a black season is coming in the world of GNU / Linux, where forks and discontent will break out even in arid deserts.
But let's get down to business. In the Manjaro forum they have published, as I said before, some isos that OpenRC makes use of. And for those who are afraid to install these versions, I leave the video on how to do it.
Download ISOs with OpenRC
The first ISO we will see is the version NetInstall. This ISO has the following characteristics:
- Based on the Manjaro-Net profile (does not have any Desktop Environment pre-installed)
- Based on the Testing branch.
- Only free drivers
- Use Linux kernel 3.14 series
- Doesn't use Plymouth
- It was tested in Virtualbox
The language can be selected at the beginning by pressing the F2 key. Once the boot process is finished we will find the prompt, where we will use to access:
- User: root
- Password: manjaro
To start the installation as shown in the previous video, we will write:
setup
Links to download the ISOs
manjaro-net-0.8.11-openrc-i686.iso (32 bit)
(md5sum: 80be54ecfb0360b2a8e544344f72113c)
manjaro-net-0.8.11-openrc-x86_64.iso (64 bit)
(md5sum: ef205f70f3b3428545fdf1420db10b74)
Post-Installation Instructions
At Manjaro forum They offer us some data for Post-Installation:
We add the openrc-eudev repository following these instructions.
1) We add the following at the end of /etc/pacman.conf
[openrc-eudev] SigLevel = Optional TrustAll Server = http://downloads.sourceforge.net/project/mefiles/Manjaro/$repo/$arch
We add and import the keys:
sudo pacman-key -r 518B147D sudo pacman-key --lsign-key 518B147D
2) We update the system
sudo pacman -Syu
3) We install our preferred desktop environment, the example uses lxde
sudo pacman -S lxde
Information on installing Desktop Environments can be found in the wiki.
4) We install a Session Manager:
sudo pacman -S lxdm-consolekit
5) We install some packages like the applet for networkmanager
sudo pacman -S network-manager-applet
6) We restart the system
sudo reboot
I think it goes without saying that for this we need to be connected to the Internet through a cable. If we use WiFi, you can see how to do it in this link.
Manajaro ISOs with OpenRC and OpenBox
In the case of the Openbox ISO, some things have to be taken into account:
- The main objective is to make the installation process easier and allow set up in a print shop network (using wicd) and the partitioning using GParted optionally.
- Configuration includes Openbox WM, LXTerminal, PCMan and NetSurf Web Browser (to search information in the wiki o google), etc.
- Use the console installer.
Links to download the ISOs with OpenRC:
manjaro-openbox-openrc-2014-11-13-i686.iso (32 bit)
(md5sum: 9be7e75c75ab296f955a3396386c4764)
manjaro-openbox-openrc-2014-11-13-x86_64.iso (64 bit)
(md5sum: 07fd57df022118dfc9e2794a0ca3d26e)
Manjaro XFCE ISO with OpenRC
Only experimentally and for 64 bits, there is also an ISO with XFCE:
manjaro-xfce-openrc-2014-11-14-x86_64.iso (64 bit)
(md5sum: e132f294f2ffd99c6cbc371d1e7a6d72)
You are right, the systemd issue begins to give off a certain whiff since OpenRC is the natural successor to the current init. We'll see where this story ends.
"Although for the end user (or for many) it does not represent anything super relevant"
I think the same, it is not relevant because as users it has not affected us in the operation of the OS itself.
In fact the only big one (debian), has given news of "scandal" on the subject, and although they say that there are other reasons all related systemd (and should not).
The other big distros, they have not made a problem (or at least manifested with spades and torches), Fedora, Ubuntu and OpenSUSE.
It gives me the impression that it is a fight between programmers, since for example opensuse 13.2 has a good acceptance / criticism and nobody in the reviews speaks of systemd (even if it is to establish debate),
Now why all the fuss of going from systemd to OpenRC, if in the end it doesn't affect them.
Personally, the systemd thing makes me uneasy, I have insecurity, good post.
In Fedora there was a certain debate about systemd when it was decided to put it as init, there were some detractors to the system, mainly because they did not agree to use it as init by default because it was very fresh and had many flaws, however, the Most of the major devs are on the core development team and were related to systemd, so the substitution of Upstart for systemd was a sign of some imposition, in addition to the issue that Upstart was an Ubuntu development and has a CLA quite frowned upon, which in the end helped everyone to accept systemd without question. OpenRC was out of the question at the time, as it lacked many of the features that it has now, including parallelization and cgroup support.
Great news! a binary distro that will release openrc… .it's like a godsend.
It is the path that archlinux should have taken from the beginning, I remember when I had to endorse archlinux to go to systemd. Now I have the possibility to retest a binary distro with openrc + eudev which is exactly what I use here in gentoo.
Thank you people from Manjaro !!!
# eix -Ic openrc
[I] sys-apps / openrc (0.13.6@24/11/14): OpenRC manages the services, startup and shutdown of a host
# eix -Ic eudev
[I] sys-fs / eudev (2.1.1@31/10/14): Linux dynamic and persistent device naming support (aka userspace devfs)
Thanks for the info, elav!
I share your opinion regarding systemd and I am also concerned about the drift that Linux has taken since the appearance of this new init. If Wheezy gets too old before the Debian fork arrives, I'll think about giving Manjaro OpenRC a try, since I don't have the free time to prepare a Gentoo system (I have valued doing so but definitely the compilation time of Gentoo is too extensive for my personal situation).
Regards!
Elav you can describe in less than 10 words for a user who does not understand the "controversy" too much, a while ago there are several articles on the blog that are being very technical, and they do not finish explaining the context for the "uninitiated" ... ever They told me that regardless of the technical, an explanation must be understood even by your grandmother to be good.
Actually in fedora, a long time ago, the problem was becoming insufferable, so much so that several desktop users were thinking of switching to centos, to bypass the problem
I sign up for that request.
Systemd works well for me. What is the problem that causes so much movement?
Let's say I don't know.
systemd is the program in charge of starting the system but the developers of this decided to extend it and now not only handles the startup but also things like cron (program to run programs automatically), the network, the system logs which by the way are binaries, among other things
Many do not look favorably on such an abrupt change, especially because it is new software therefore with many more bugs than programs that had worked all their lives, in addition to generating dependencies when programming and for example gnome is increasingly linked to this system. Making it less portable to other unix platforms.
I do not know if my other comment did not pass the moderation but it said that I liked systemd but they should not let it monopolize all the distributions and leave alternatives as has always been done in Linux for those who have different needs.
I fail to say that before the program in charge of starting the system at boot was system v, which had been going for a long time until it was replaced in most distributions by systemd xD.
To what @daryo says I add the following (which is my opinion also):
I've always liked the Unix philosophy where a program does only one thing, but does it well. When Systemd wants to control everything that @daryo told you, I have a little doubt and what would happen if Systemd is somehow compromised? Well, it would possibly drag with it everything it controls.
To that I add (and perhaps this is more out of habit), that I have always liked that my system logs are pure text files, but with Systemd everything is binary, and commands such as:
cat log.txto
tailf log.txtWhere we could use other options like GREP to filter certain content, but Systemd uses a named command journalctl.
In addition to the aforementioned, I must say that being RedHat the main exponent behind Systemd, I get an alert that I cannot turn off. Maybe I'm wrong, but this does not look good .. And I keep wondering what need is there to control the boot, the cron, the network and how much service exists? What do they mean by that?
Thanks to your comment and what I have been investigating, I can confirm your suspicions, that alert is correct Broder.
You see, I have been reading about TCP Stealth, it is a German thesis where they accuse Red Hat of facilitating industrial espionage to the listening systems of the 5 eyes:
I already wrote about this, if you have the necessary talent, I know you have it, you can reach your own conclusions:
https://gnunet.org/sites/default/files/ma_kirsch_2014_0.pdf
http://heise.de/ct/artikel/GCHQ-NSA-El-programa-HACIENDA-2293098.html#TCP Stealth
Just to complement your nice comment @elav, systemd is NIH so high, that it now claims to control the following:
1.- Management of internet connections with IPv4 and IPv6, using systemd-networkd and systemd-nspawn.
2.- DNS management through an internal DNS cache, systemd-resolved.
3.- Multicast DNS management in internal networks, using systemd-networkd.
4.- Management of TTY terminals in Linux, using systemd-consoled. (Goodbye KMScon?)
5.- Management of sessions and privileges through logind.
6.- Coredump control, using binary files, and skipping the kernel directives.
7.- Control of logs, using files in binary, and skipping the kernel directives.
8.- Control of ACPI events using logind. (Systemd-212 added several headaches to Nvidia devs with various bugs that rendered the system useless)
9.- PPPoE support for networkd, a job that is still in progress.
10.- Support for DHCP in client and server. (What do they do with that? No idea)
11.- Support for systems with factory reset, which by the way is closely linked to BTRFS (Don't be surprised if BTRFS later becomes a dependency on systemd, good Lennart loves it)
12 ..- Support for virtualized containers (Xen and KVM mainly)
13.- Support for device handling and initialization (what udev does)
14.- Handling of disk encryption systems.
15.- Loading firmware and kernel modules.
16.- Handling of hostname (it creates up to a unique identifier of your PC), premises, time, NTP synchronization, sysctl (kernel control variables), and even the random number generator (Very WTF this, and it raises a lot of suspicion )
17.- Handling of temporary file systems.
In short long list, are the things that I know that systemd does, if anyone knows any more than say it :).
PS: systemd no longer offers support for LSB and SysV scripts since systemd-214, so I don't know how true its "legacy" support is now or how standards-compliant it is. I say is LSB still the standard in Linux, or am I wrong?
Thanks for letting me know, I was thinking of going to BTRFS but knowing that Lennart likes him, you may know that he must be horrible and spy on the NSA-IBM
There is little space to summarize and explain so much ... it is a giant Trojan horse, which they do not even try to hide. What does a startup system do by putting network services, dhcp dns and even I think avahi ... into systemd? Power of decision is lost by not being able to manage services
that are not wanted and that they do not come to me that can be deactivated, I do not want them in the systemd package!
In OpenRC, one is the one who decides what things are started in each runlevel, some services have dependencies on other services but they are very few and are listed ... while in systemd anything does whatever it wants at the moment He feels like it… to gain about 5 seconds on the boot and be quick on the shutdown.
Systemd is so complex that it is impossible to know what it does, you have to resign yourself to thinking that it is your master and does not do anything wrong to you.
Systemd breaks the concept that things should be easy and understandable in terms of daemons or services and runlevels, no one who uses systemd fully knows what happens in their services at all times.
Systemd does not allow using syslog-ng natively, they have set journald to step on it and it does not let it work, that is, or do you use journald or naninga !. The system log is fundamental for the security and auditing of what happened and is happening with the local and remote connections, but journald uses a binary format that only jornalctl can see it…. Very often journald gets corrupted "mysteriously" his binary file and as he sees it corrupted he deletes it once and starts with a new one, forgetting all the logs that already existed.
I can go on for hours, but the worst problem is that Lennart does not give a ball to those who report those errors and as far as I read, he does not accept patches from anyone.
I thought that when they got into systemd, they would report bugs and patches, which the systemd would have to accept ... but I honestly believe that Lennart and RedHat have another plan for the rest of the distros .... as I said before, HORSE OF TROYA from RedHat.
Honestly for me systemd is not fixable, the idea behind its design is perversely bad, it is better to start a bootable system from scratch than to try to fix that frankestein.
AMEN!! @anonymous..
I've been using systemd (Fedora) for like a couple of years and I've come to this:
The issue smells weird, as more things add more disable / redirect.
The journald I have directed to rsyslog directly. Some binary log of yours has already been broken.
From dns I use bind, if they integrate it into systemd I will continue to use it the same, even if I have to modify everything.
I use XFCE so it saves me a lot of what gnome wants to integrate.
It is like an elephant in a china shop.
True; not even they know what to call it. We go out to update daily, correcting bugs and other crap. It is a subject that makes me quite angry; but not only because of the fact that SystemD is a sovereign shit; if not how they have done it.
It is clear that in the Linux world, there are several companies that try to control everything; see Canonical, RedHat and Gnome, (even Miguel de Icaza himself has left Gnome).
If I use Linux it is because of who I control it and that is the basis and philosophy of it; In order not to know what it is doing, I mount machines with W Server already running.
What I'm sorry for is that Debian has succumbed. In fact, the possibility of creating a parallel fork without SystemD is being considered.
Let's hope that the thing does not go to more; or I see myself migrating all my machines to BSD.
@ anonymous, comment piece man, you can't be more right.
systemd is a crazy thing that has no explanation in many things, the truth causes a lot of suspicion in everything it does and does not allow other tools to do it, the truth is I don't know how the Debian people allow themselves to put this, but in the end they already made that decision , and for the first time in many years I have stopped using Debian as the main OS, and will continue until systemd exits Debian for a more transparent option.
In short. SystemD sucks.
It stores logs in binary format, it is executed as the parent process of all the others, (Pid 1), with which if somebody breaks, the system becomes unrecoverable; It goes against everything Linux stands for, that is, plain text files, (what the hell is that binary files ??, plain text files! Like all the life of God.)
Come on, that's crap. I don't like anything at all.
But thanks to companies like Canonical, Gnome and Red Hat; we are going to eat it with potatoes.
That if, while there are other options; I will not use it, neither on the servers I administer, nor on my personal machines.
This is already becoming a branch of the Redmond company.
I don't mean to defend anyone, but I remember well that canonical was totally against systemd in favor of upstart. when debian gave in to systemd it ended up dragging to ubuntu.
In addition, these bugs can compromise the security of the system and the stability of a server, for example, that is why the ones who complain the most about these things are the sys admin.
And what about Mageia, it is incredible that a KDE can run on 512 MB of ram, impeccable.
http://mirror.cedia.org.ec/mageia/iso/cauldron/
a few questions; How easy is it to manage the services in OpenRC? And how easy is it to install using it by default in a Manjaro installation with systemd? what I like about systemd is that with the simple command systemctl enable (service) .service or a systemctl disable (service) .service I can manage my services easily, IF I am interested in knowing about OpenRC and especially if it smells a little strange all this from systemd, by the way; I am a novell user
By the way; It says I'm on Windows because I'm using user agent overrider
OpenRC is very easy to handle, I give you an example with the cupsd printing service.
To start it.
# rc-service cupsd start
* Starting cupsd .. [ok]
To stop it.
# rc-service cupsd stop
* Stopping cupsd… [ok]
To restart it.
# rc-service cupsd restart
* Stopping cupsd… [ok]
* Starting cupsd .. [ok]
To put it to start in the default runlevel.
# rc-update add cupsd default
* service cupsd added to runlevel default [ok]
To remove it from the default runlevel.
# rc-update from cupsd default
* service cupsd removed from runlevel default [ok]
To see the status of all services in all runlevels.
# rc-status -a
To see the status of a runlevel, in this example default.
# rc-status default
Here in gentoo, OpenRC is the default startup system and will remain so forever, we have systemd in portage for suicide bombers, who luckily are few….
To replace journald, we use syslog-ng and logrotate, here in gentoo the system log goes out through the virtual console vt12 that is control + alt + F12, or you can see it continuously in any graphical terminal as root user with:
# tailf / var / log / messages
And to install it on my Manjaro?
I say; I'm not going to lose all the files and my beautiful XFCE just for switching to OpenRC 🙂
Ready; I installed it by using sudo pacman -S manjaro-openrc bluez-openrc (the latter because I have bluetooth)
Now my problem is that XFCE4 power manager doesn't work with upower-pm-utils 🙁 and I don't have the typical options of suspend and Hibernate
OpenRC is very simple, managing services is a piece of cake, just to give an example:
Enable a service: rc-update add cronie default
Start a service: /etc/init.d/cronie start or rc-config start cronie
Stop a service: /etc/init.d/cronie stop or rc-config stop cronie
Simple and not really complex.
@elav what lies ahead is for the long haul, ranging from sandstorms, trolls rain, bulk forks, dev group splits, and many wondering if migrating to BSD is a better option than getting stuck in systemd, because yes.
Personally, I applaud this initiative of Manjaro, it is an option for those who do not want to stay with systemd, something that I like, at the moment I am in Gentoo and I like it, I feel comfortable with the freedom it gives me, but now It has crossed my mind several times to make the change to FreeBSD, and I may make the jump this month, it all depends on my time and ordering certain things to carry out the migration successfully.
The arguments against systemd have long been refuted. Read a little more please.
http://0pointer.de/blog/projects/the-biggest-myths.html
http://diegocg.blogspot.mx/2014/02/la-sombras-de-sysyinit.html
http://diegocg.blogspot.mx/2014/02/por-que-kdbus.html
None of that refutes the reality of systemd, Lennart is very good at avoiding things and responsibilities, I recommend that instead of just reading articles, read the systemd code or at least read the devel list of systemd, you will find out about things that refute what those three articles say that happened, and more supportive of systemd detractors.
His argument is to establish that there is knowledge that refutes what I showed, but never presents the evidence, so I cannot trust its existence.
https://lists.debian.org/debian-ctte/2013/12/msg00234.html
@pamp my argument is a bit more support because I explained it above in comment 25 of this same entry, and I have exposed it in many other entries regarding systemd, in addition to exposing it in the Debian irc and the list of this distribution, also my invitation is that you create your own opinions and for this you just have to read a little the devel list of systemd. Also just to pique your curiosity I give you this link in which they clearly say that systemd-214 no longer offers support for SysV and LSB scripts, with the excuse of "code cleanup."
http://lists.freedesktop.org/archives/systemd-devel/2014-June/019925.html
Now tell me: Where is the support for the LSB standard that is supposed to have been created in order to create a common base for all distros? Because let me tell you something, nothing else on his first link Lennart plumps, brags and fills his mouth saying that systemd supports the use of SysV and LSB scripts, when the truth is that support is dropped and replaced by a generator of init-files, which by the way, has several bugs and in the end there is no choice but to make a complete init-file.
Greetings.
The opinions, it's like the ass, we all have one.
What this man says may go very well for him, but it is not my case. And the opinion of a man who writes on a web portal is not that it is the word of God. It is your opinion, period.
So "refuted", nothing.
The good thing we have left is that we can use whatever we really want; without trying to be "Taliban" and impose our criteria on others.
For me SystemD is a real shit. And there are people who love it. Well welcome!
Neither my opinion is good nor that of those who do not think like me is shit; they are simply different.
This is what distinguishes us from other operating systems; we can choose.
Let's not get into useless fights that lead nowhere.
@Titus
You couldn't have said it better ... amen.
You have to be blind not to realize the pervesity that systemd drives to cover everything, stepping on, covering and displacing projects that work perfectly, replacing them with versions that never reach or become stable, even there is no compatibility between cores and more than two back versions of systemd.
Debian seems to have gotten the earthquake and managed to wake up, I just hope they lean towards eudev and openrc, so the development of gentoo debian manjaro and some others that use openrc would be unified, which would improve it a lot in a short time, winning the whole community.
I second your words.
There are people who quote other people (the opinions that interest them, generally), and wield them as evidence.
For my part, I don't have an opinion on systemd. I don't know if it's technically better than upstart or openrc, but what does seem clear is that the possibility of sysvinit is ruled out by ALL distros, with Debian being the only one that still kept it in Wheezy due to its policy. But the next stable Debian, Jessie, was going to be a Debian without sysvinit.
What is clear is that ethically it is 100% free software; As for its technical part, I have neither studied the code nor compared it with its alternatives, so I have no reasoned opinion. But even current Ubuntu uses parts of systemd even though they still have upstart, and I doubt they did because Canonical is "bought" by Red Hat.
Systemd is not "evil", please, we are not fighting Skynet (Terminator), or HAL9000 ("2001 space odyssey"), nor is it the dark side of the Force that seeks to dominate the Jedi. Nor is it that by settling into a team, it takes over everything and makes even the food items in the pantry disappear.
And that that "it moves projects that work perfectly" (comment 52), I have had problems with a home NFS network in the computers that access the server because the shutdown process of the client computer disconnects the network before dismounting the NFS system, and the shutdown would freeze, the only solution being to press the on / off button to turn it off by force (bug reported by various users); I had to create a script that unmounts the NFS files to run before shutting down the client machine. On the other hand, the NFS server computer connects via wifi, and from time to time the connection is lost: I don't know if the problem is the network-manager or it is in dhcpd, or where.
I'm not saying that those problems go away with systemd; I ignore it, because I have not used it. It is just a sample that to say that the projects that systemd replaces work perfectly is an exaggeration.
One thing is an opinion and another is an argument, certainly the first is very varied as you say @Tito, but the second is something more concise and focused, it is not something that can be so easily manipulated, at least, not in the case of free software, where we have the code at our fingertips to review.
@pamp tells us that the arguments shown have been refuted for a long time, and as a first test he brings us up to date with Lennart's opinions (not arguments). But what this guy says in his comments is one thing (numbers 4 and 8 are just to die for laughing), and what he does in systemd code is another. An attitude that I have seen repeatedly in Lennart since I started developing things like Avahi and Pulseaudio, and that can simply be corroborated by reading the devel lists and the bugs reports of both pieces of software.
@ Dah65 certainly many people wield evidence using the opinions of third parties, a bad habit for those who cannot investigate the issues themselves to have their own and personal opinion, and even create valid arguments with which to participate in a constructive discussion .
In my case, I keep abreast of the changes in systemd thanks to the devel list, although I don't like the tool, I totally dislike it, but I don't stop reading about it at the user and technical level, and the reason For this it is very simple, if I have to attend a client that uses said init, I know what I should do and how to attend to any situation.
Now about what the services run without problems, that is a fallacy, there are many SysV scripts with problems, and the same happens in systemd, but at least when you report an error in a SysV they are fixed or you can do it in a simple way as you have commented , in systemd, after making a bug report you can find a WONTFIX or CLOSED, thanks to Lennart or Kay, as the case may be and I am not exaggerating when saying this, a sample here:
https://bugzilla.redhat.com/show_bug.cgi?id=753882
Read comment 48, you have no loss. Clement's 53 is another that has no loss, especially for its archaic but functional solution to the problem that Lennart does not want to solve and that by the way was reported in 2011.
Those "myths" who established them? Some are removed from the gallery as "systemd is not portable for no reason." It is completely true that it is not portable (and he admits it saying that it is very customized for Linux)
It assumes fallacies, such as the assumption that BSD is not interested (the BSD guys say the opposite: "Jordan Hubbard - FreeBSD: The Next 10 Years (MeetBSD 2014)"), even if it was portable they would not adopt it and things like that ( myth 13,14,15).
If Poettering's intention is for us to start rewriting scripts, exclusive to your system (http://0pointer.de/blog/projects/systemd-for-admins-3.html) we are going to go wrong. In principle a classic init script doesn't care where you are going. Minimal modifications are made to work on GNU, UNIX or BSD. Well that was until now (unless OpenRC is used). Anyway, I think things like this will produce a schism between Linux for desktop and servers. Ubuntu and derivative users will only see the changes at the end of next year.
@ Dah65
Well, since you say that systemd is not perversity personified, then tell me why they don't put in the Makefile options to disable all its modules at compile time, so that those of us who don't like to have those "optional modules" that step on other packages, so we can compile them and create our own versions of systemd capped!
Do you know why they don't? Because its form of development is called forced imposition and as 95% of users do not have NPI, they take advantage of the default, we declined it for all of you.
That's how free or opensource software or whatever they want to call it doesn't work, now it makes me laugh, because with the new fork to Debian many people come out to think that it is a waste of strength and I keep wondering how difficult it was to put additional compilation options to the Makefile?
The subject does not give for more, this is like wanting to mix water with oil, that is why there will be endless forks in each development where there is an imposition of a few for all the rest.
@mario is exactly what you say. Jordan Hubbard has also come to his senses that the BSD init needs to be updated not only to adapt to new technologies but also to support new features that are now possible, but he bypasses the concept that systemd now has of how they should be done. things, and they simplify it to the philosophy that has always prevailed in UNIX, "Make a program that does one thing and does it well", and that in an init is of utmost importance, since we are not talking about one more demon, we are speaking of the init of an operating system, in addition to being a security measure, compared to what many specialists are already beginning to rant about systemd, and it is demonstrable, systemd looks a lot like svchosts.exe from Windows, doing from the init of services to network control among many other things.
Guys, it's really scary.
Is it very complicated to remove from ArchLinux ????
I'm going to look for information but I don't dare to touch that kind of thing lest I screw up and lose my system.
: )
https://abchk1234.wordpress.com/2014/06/03/using-openrc-on-arch-archbang-manjaro-linux/
From the many comments I have read, SYSTEMD is a true TROYAN HORSE….
This means save who can? There is little information in Spanish -about the desktop configuration in FreeBSD and getting the system ready to be used.
Poor systemd, let it be. xD
This hate systemd will not be viral ???? Arch has gone great for me ... if it's true that it's covering more, I don't know if it's good or bad! but perhaps there are already vulnerabilities to take control or some virus that destroys the system due to this ... if it is stable and safe I do not see the problem ... anyway I will see if I have time and study the subject and do some tests with openrc
not that stable. and it is much more insecure than system v. For a desktop user like many of us (me) it does not represent a problem either, a faster boot works well and I don't usually read the logs so it does not matter how clear they are or whether it is in binary format.
I have a theory linux will grow on desktops (and governments) and lose ground on servers (instead of taking OS like freebsd)
On the esdebian Wiki they publish how to install SysVinit on Debian Jessie. http://www.esdebian.org/wiki/sysvinit
Reading about security, I find out that on the intel side, there are motherboards with chipsets, generally in the northbridge, they implement something called AMR Intel Active Management Technology .... interesting, luckily I don't have intel, but I'll start looking for it AMD side there is no such thing.
They imagine a combination of intel + AMR + systemd, God forbid.
https://en.wikipedia.org/wiki/Intel_AMT_versions
No wonder Stallman's paranoid cries out for free bios.
Quiet; Intel abandons the manufacture of motherboards.
http://www.infoworld.com/article/2612845/computer-hardware/intel-refocuses-and-exits-motherboard-business.html
First of all, I don't use systemd because it's not built into Kubuntu yet (I'm with Netrunner 14, derived from Kubuntu 14.04).
Having clarified this, several things must be specified:
1- systemd is being adopted by developers / packagers of many different distros (Debian, openSUSE, Arch, Fedora…), but now it turns out that the readers of this blog know more than they do about the advantages and disadvantages of systemd.
2- systemd is free software, whose code can be read (and understood), by those who have time and knowledge (those developers / packagers I was talking about before). If you hide back doors, they would be discovered. How many of the readers use a proprietary firmware or driver, the code of which you have not read and cannot read? I think more sensible to be afraid of that than not of systemd.
3- We all work with binary packages, because when I download a .deb from the repositories to install it, I am not downloading a plain text file. So that argument is quite paradoxical.
4- In GNU / Linux there are already programs that do many things: the same kernel, which increasingly integrates more drivers, and even proprietary firmware (better to put a back door in the closed firmware than in a program whose code is published). There is also Xorg, which not only handles the graphical server but also the keyboard, the mouse, and other things; Nobody says that Xorg "betrays" the UNIX philosophy for that, they want to retire him because he has already been overtaken by other projects.
5- "Linux is choice", of course, but it is the freedom to choose if I want to read the code, alter it, distribute it, etc. Not that distros are required to give all choices (all processor architectures, all desktop environments, all package formats, etc)
6- For those who are thinking of switching to a BSD, I remember reading news that in some BSD systems the American NSA had already put its claws. If this news was correct, I do not know because I have not followed the topic. But it is ironic that I flee from something "because Red Hat is behind and maybe ..." to get into something that "maybe the NSA is behind ...."
In addition to using GNU / Linux, BSD, Windows, or whatever you want us to use, we can also use our logic and ability to reason
In other words, the readers of this blog, being only readers, do not have the ability to realize if something is good or not, because we must be guided by the good judgment, knowledge and experience of packagers and developers. Is that it?
It is true, it is Free Software, and if something strange appears, the super people you talked about before and that we must trust will be able to notice and announce it, or perhaps not, because maybe as they are people they will be tempted to shut up in return Of something.
When you download a .deb, all you are doing is downloading a compressed file, which you can unzip and therefore see what is inside and is possible, which is where the binary is inside. 😉
I don't know who are the users who are going to flee from Linux to go to BSD, but for example, I would not have to leave Linux, I would only have to leave a distribution that puts Systemd behind you yes or yes.
In short, those of us who comment, read and use GNU / Linux in this blog do not reason. That's what you want to say? Anyway I'll tell you from my personal experience, and my reasoning (be it logical or not):
Systemd is shit stuck on a stick. I have read that there are other Inits that start much faster and therefore do not have to control DNS, RED, CRON and everything else that Systemd wants to control. Maybe for an end user, who only cares about turning on the computer, opening a browser and sending emails, it does not matter if they use Systemd or Systemx, but for those of us who manage servers it is a pain in the ass. And I ask you the same question that I always ask what happens if Systemd is compromised and goes to hell? Are we left with no RED, no CRON, no DNS, no Init and everything else it does? There I leave it for you.
And beware, I tell you all this without acrimony. That said, welcome to these parts.
Thanks for the welcome.
Responding without acrimony, I clarify that I neither develop systemd nor get paid to promote it. And that it does not affect me at all whether other people use it or not, it is their decision.
But what I see with this matter seems, at times, a hysteria, and I read opinions of people who, without having studied the code or having used it, label it as garbage, imposition, treason, and I don't know how many other things. It reminds me of a situation I experienced a few days ago, when a person who acknowledged that he had never installed Windows or knew how to partition a hard drive began to say that Linux was very difficult ... without ever having tried it, and also having Android on his smartphone.
Have you compared systemd with sysvinit, with upstart and with openrc? Great, you can make a decision based on your own experience. It is the best, because you also know that the distro that works on one computer can be worthwhile on a different one, and that is why those of us who have some experience in GNU / Linux say that the best distro is the one with which the user feels comfortable. taste.
1- «In other words, the readers of this blog, because they are only readers, do not have the ability to realize whether something is good or not, because we must be guided by the good judgment, knowledge and experience of packagers and developers»
I have been a reader of this blog for quite some time (you will see my comments in old news), so I am included in the pack. And the answer is no: being a reader of this or any blog does not enable me (at least for me), to judge the good or bad of a software that I do not know. I can read what others are saying, and in this case there are positions both for and against systemd; in fact, every time the topic is raised in Phoronix there is much debate, but even there argued comments are scarce. I'm referring to arguments like "when systemd calls process X an infinite loop occurs, rendering the system unusable."
And the truth is that by using a distro or a different one, you are being guided by the judgment, knowledge and experience of packagers and developers. The use of any OS or program implies in part relying on the judgment and experience of others; for example, with Linux you accept the decision to use a monolithic kernel instead of using a microkernel like Hurd. That decision was Linus Torvalds's, and you accept it by using his core.
2- «It is true, it is Free Software, and if something strange appears, the super people you spoke about before and whom we must trust will be able to notice and announce it, or perhaps not, because maybe as they are people they will feel tempted to shut up in exchange for something. "
Well, suspicious, why trust Linus Torvalds and Richard Stallman and the GNU project? I have not looked at the code of their programs, so maybe they are deceiving me.
3 - «And I ask you the same question that I always ask, what will happen if Systemd is compromised and goes to hell? Are we left with no RED, no CRON, no DNS, no Init and everything else it does? I'll leave it there. »
What if OpenRC is compromised in some way? Or Upstart? Or the kernel? It happened to me, after a "normal" update in Debian Testing, I ran out of grub, I couldn't enter Debian or Windows, and at that time my ignorance meant that I only had the option of reinstalling.
4- «In short, those of us who comment, read and use GNU / Linux on this blog do not reason. That's what you want to say?"
No, I don't mean that; I do not intend to generalize from a particular, concrete situation, to the totality of the behavior of one, or a thousand, people. But I do believe that in the case of systemd it is spoken many times without doing an objective and serene analysis; it also happened with Wayland-Mir, many unsubstantiated claims being made, both against Wayland and Canonical.
Also, I repeat that I read and comment on this blog (as in others), and that I use GNU / Linux.
And I also repeat what I said before: let's use our brains, analyze what we hear and read, take different perspectives to try to refute both A and non-A, and if possible, let's get our own experience to base our conclusions on facts. And then let's use whatever feels right to us.
umm .. well that being compromised is a hypothesis is like everything .. my question has already passed? .. maybe bugs are not found in all software and they are corrected if there are bugs in systemd they correct it and as any program can have its bugs .. the problem is not that it can fail, it is if you want it to do or control what it is doing but not on the assumption that it can fail anything can fail in a moment ... I'm not a fan of systemd at all, it's just my opinion .
A bug can occur on a user's computer and nothing may happen, but on a server things are very, very different.
@waco certainly if you get bugs in a software you should correct them. The problem is that systemd has a lot of old bugs (some dating back to 2010 and serious) and they are still not fixed today, or simply downplayed, or simply marked by Lennart as CLOSED or WONTFIX.
your comment is very successful! We can not all fall for systemd because it is fashionable and has been created as a smear campaign to this ... every change has a rejection.
I respond to your arguments:
1.- Serious and inquiring users, and developers alike know the advantages and disadvantages of adopting systemd in any development and work environment, the weaknesses and strengths of systemd do not change due to having one or another perspective.
2.- Certainly systemd is free software and can be audited. The problem is not that it has hidden back doors, the problem is that it does things that an init should not do (network control, dns, TTY consoles, etc.), that it has a lot of services that are supposed to others, that it does things in a completely different way from how they are expected to be done, which breaks the rules of the Linux kernel itself (coredump), that many of its developers care very little about solving structural problems that systemd has (coredump and debug are among the most serious yet not solved).
3.- One thing is to download a binary that turns out to be a program whose CONFIGURATION and LOGS are still in plain text, and another thing is to download a binary whose CONFIGURATION and other information is stored in binary and is only accessible through specific tools, This is where things change. A binary log does not offer security (if you really want security, encrypt the partition with AES-256), it is just a black box from which you do not know anything about what is happening and it lends itself to many things, for example: Imagine that you have a Trojan that exploits a systemd vulnerability and through it obtains full access to the system including the log service and privilege escalation. Isn't that a serious problem? Wouldn't the binary logs handled directly by systemd turn against you by being inauditable without getting to the point that they have already been unknowingly modified? There is the point and difference between a program and a configuration file / logs / dumps in binary.
4.- The kernel is a piece of software designed in that sense, it is designed from the beginning to control everything on your PC, not an init. An init is dedicated only to making your system lift the kernel and be usable, because it is the first thing to start and the last thing to finish. That is why it is called init (initialization) because it only starts the system and does nothing else, and the reason for this is very simple, the init must be the most stable and perfect piece of software possible, to avoid that for some reason This ends up breaking the whole system, it is about stability and security. Xorg, it is another voice, it does many things it is true, but nothing so risky as to leave you with a completely unusable system, and also its configuration is still done in simple plain text files.
5.- Certainly distros are not obliged to offer freedom in a broad sense, and it is because of this that the current tirade is presented. But, we are users and the community, and many of us simply do not agree with the implementation of this system, that is why we make our voice reach, whether they listen to it or not, it is a matter of those who develop the distro, and their decision will have an impact on those who decide to use their distro or not, and that, clearly can lead to the failure of several distros depending on how things are going and an example now is Debian and its Devuan fork.
6.- The BSD news is because of what happened in OpenSSH and in the OpenBSD IP stack, a back door that by the way affected not only BSD but also Linux (in the case of OpenSSH), and that was fixed. The situation is attributed to BSD, because it is BSD (Theo de Raadt in OpenBSD) who are in charge of the development of this tool (OpenSSH) and the situation arose because certain developers who are no longer active in the project planted the back door. The situation was resolved and the pertinent measures to be taken were declared in case this situation could be affecting those who made use of the software. Now: Can this situation occur in systemd? The answer is simple, and the result is catastrophic, since systemd handles escalation of privileges among many other things, a backdoor in systemd means total access to the system, something that did not happen with the backdoors mentioned in BSD.
They return the Debian fork without systemd already has a web page. It seems that the project is going and very seriously. https://devuan.org/
Updated ISOs and some new uploads.
https://forum.manjaro.org/index.php?board=50.0
The installer is not very clear, I cannot follow their steps, especially in the part of the partitions, I do not know why they insist on these confusing things.
There is something that catches my attention about the netinstall with Openrc, somewhere in the installation I keep seeing the message that you are configuring systemd, will they really be free of systemd or its use?
Hello Manuel, I also observe the same thing during the installation, it must be a matter of the installer because what there is no doubt is that systemd is not installed, confirm in the terminal like this: pacman -Qs openrc
regards
Hello keos, first of all I apologize for not having answered before. I appreciate your answer, I am happy to know that Manjaro offers this option; as soon as Ubuntu Precise support ends (or maybe sooner) I will install it. Regards.
Good post
I'm going to wait in Manjaro with Systemd while the OpenRC version matures a little more, I want to get out of systemd… (I sweat it)