SlackwareThis Forum is for the discussion of Slackware Linux.
Notices
Welcome to LinuxQuestions.org, a friendly and active Linux Community.
You are currently viewing LQ as a guest. By joining our community you will have the ability to post topics, receive our newsletter, use the advanced search, subscribe to threads and access many other special features. Registration is quick, simple and absolutely free. Join our community today!
Note that registered members see fewer ads, and ContentLink is completely disabled once you log in.
If you have any problems with the registration process or your account login, please contact us. If you need to reset your password, click here.
Having a problem logging in? Please visit this page to clear all LQ-related cookies.
Get a virtual cloud desktop with the Linux distro that you want in less than five minutes with Shells! With over 10 pre-installed distros to choose from, the worry-free installation life is here! Whether you are a digital nomad or just looking for flexibility, Shells can put your Linux machine on the device that you want to use.
Exclusive for LQ members, get up to 45% off per month. Click here for more info.
I would prefer Patrick stick to bsd-style sysvinit. Honestly, I have enough of my system ran over cronjobs and other alternative real methods that to me change is pointless. Don't get me wrong my time with Runit was eye opening, but honestly if you do things the right way, you don't need to change things majorly, if at all. Cron mixed in with init as well as starting services with desktops rather than boot time, shaves off great deals of time and does more work than you realize.
Cron is the best service supervisor you can get. It's not automated, but it can be used to alert you the administrator to get off your ass and do your job if something happens. Have cron check services at regular intervals and file a log report out via email or intranet messaging services, or have it run a script to perform an if/else check using grep and other tools to restart a service if it's down is just beautiful. Cron is like the king of Swiss Army knives when it comes to getting work timed and scheduled, een on boot. To be honest, you can use cron to start services in parallel on boot once it comes up, and then use scripts to check the execution state and perform a trigger if it's down to relaunch it.
If you learn your system, you'll learn you can do more with less and using existing tools to do everything other packages try to consolidate, and nothing beats effective shell scripting... Nothing!
Or he had no choice, because systemd permeated the system to the point that the system could not run without it. (There are usually at least two scenarios for every situation.)
You could say the same thing about any major component in the software ecosystem. The only way that a distro can do to guarantee total immunity to upstream change is to become the upstream. The arguments on this have been had. There are a group of people who, for quite legitimate reasons, prefer to administer a traditional human readable system. There are others that have to manage literally thousands of servers in scenarios that we could not have predicted when the 'nixes were conceived-for these people systemd works.there is literally no point to rep
eatedly raising the same points over and over- every point has been addressed in the discussion. The discussion that we do need to have is about what sort of community we want to have and how we want others to see us. The sort of vitriolic ad hominem that is exemplified by your current signature does nothing more than make us look like a bunch of illiterate, pitchfork wielding idiots. Please, please please just drop the Lennart hate and ask your self; is the likeability or otherwise of an individual developer a useful criterion for judging the usefullness of a piece of software?
Last edited by Phorize; 10-19-2014 at 05:17 AM.
Reason: Typo
I don't hate Lennart. I just hate it when he and others like him force people to feel stupid, ignorant, and useless because things aren't cool, revolutionary, or doing new things for the first time their way rather than the UNIX way. For that, I don't hate Lennart... I wish him completely erased from history. No one should ever be allowed to make people feel less than their potential. That is a cardinal sin again others, especially in software.
Fadware has no place on UNIX and UNIX-Like systems on any level. Letting people learn to use existing tools, promote usage of those tools, and teaching new ways to use old tools in new ways with simplicity makes people feel smarter, teaches them more, and serves as a helpful guide to expanding knowledge and creating concrete foundations in UNIX and UNIX-Like systems like GNU/Linux.
You could say the same thing about any major component in the software ecosystem. The only way that a distro can do to guarantee total immunity to upstream change is to become the upstream.
It will only be an upstream issue after the thing in question is worked into the kernel?
Quote:
The arguments on this have been had. ... The discussion that we do need to have is about what sort of community we want to have and how we want others to see us.
In other words; accept it and show the world a happy face.
Quote:
Please, please please just drop the Lennart hate and ask your self; is the likeability or otherwise of an individual developer a useful criterion for judging the usefullness of a piece of software?
Read what he says and writes, and look at his other projects. There is much fuel there to not like the guy. People who read Poettering's propaganda will either despise him or mindlessly believe it. For an example of his past work, look at his wonderful Pulse Audio, which was required, because ALSA is too simple. Unlike Microsoft's audio software, which is great.
Of course, I am left wondering why this outburst was posted. When the discussion turned to how Volkerding might manage to avoid putting systemd in Slackware, I mentioned one possibility; he might not be able to. Why does that cause people to foam at the mouth and call for the infidel's head?
On that note, this will hopefully be the last post I ever make about systemd on a Linux board. Hopefully the deities will give me enough strength to avoid the temptation to get involved with any more discussions of the subject.
The thing is that systemd is a solution looking for a problem.
one of the problems is (re)booting a view thousands virtual machines as fast as possible.
some companies have such problems, and one of them decided to pay some developers to find a solution.
the outcome is systemd, you may like it or not, but as long as it has businesses value, and there is no better solution, it is here to stay
and since there are some other problems also, like the logging stuff, the devs implemented their idea of problem solutions.
and obviously some big player needed this solutions and therefore they exists.
you may like the or not, but fact is that systemd is used to run system critical stuff for a million dollar businesses, and it does this because other solutions have not been good enough or do not fulfil new requirements we have.
want to replace systemd, no problem, make a demo that targets the same problems but solves them in a better way.
than companies will invest money and force this solution.
How many of you are willing to abandon KDE,XFCE,MATE,GNOME... because of systemd.
The crap-kit was forced into Slackware because any desktop environment out there depended on it. Now all these DE are happy to replace console-kit by systemd-logind. Simply because the developers are happy to have an unified way to interact with the underlying OS and to concentrate on their main task developing desktop applications.
How many of you will do mknod /dev/... because there is no a real replacement for udev and udev is part of systemd.
If the attitude of developers is a reason to not use systemd then I wonder why we still use the kernel.
If sysvinit is good enough why are there so many alternatives.
If PAM is a SCAM why everybody uses PAM+LDAP. They could happily sync thousands user accounts between thousands workstations.
If someone is happy with his large home network of three computers why the world outside of his basement should stop moving.
I don't think the cheap talk has some place here. And I am not willing to do any boot time benchmarks because I simply don't care about them.
So then why did you try systemd ? I thought speed was the main thing here, you didn't mention anything else. Either way "much faster" doesn't mean anything.
Cron is the best service supervisor you can get. It's not automated, but it can be used to alert you the administrator to get off your ass and do your job if something happens. Have cron check services at regular intervals and file a log report out via email or intranet messaging services, or have it run a script to perform an if/else check using grep and other tools to restart a service if it's down is just beautiful. Cron is like the king of Swiss Army knives when it comes to getting work timed and scheduled, een on boot. To be honest, you can use cron to start services in parallel on boot once it comes up, and then use scripts to check the execution state and perform a trigger if it's down to relaunch it.
This paragraph is an excellent advertisement for systemd. Cron is not a service supervisor! On SysV init, if you need to restart a service when it crashes, you have a few options.
1) You can wrap the service in a script that relaunches it if it dies (using `wait $!` in bash or something more complicated in another language). This has the overhead of bash/ash/etc. and unless you're particularly clever it may require a different script for each service...yuck! You're also stuck playing around with PID files so you can properly stop the service without causing it to restart. I have done this myself but it isn't pretty and it certainly isn't ideal. It is also unlikely to work well with Slackware's existing init scripts so you may need to write a start/stop/wrapper script for every important service. Ew.
2) You can use cron to monitor a service at regular intervals and restart/take action if something is wrong. This is really ugly -- if the service is one that is expected to be up at all times, this solution is not usable (unless you want the service to be down for up to a minute, or longer if the check is performed less often). This also means that a check, using up resources, will have to run every minute -- if many checks are all running at the same time (or within seconds of each other anyway), this may cause a regular hiccup, which is unacceptable. If you decrease the frequency of these checks to mitigate that problem then there is a longer potential downtime.
3) Monitor it yourself and restart it when you notice it is down. For hobbyist servers this isn't really as bad as it sounds, but it certainly isn't an option for anything serious.
Using such a setup relies on services generally being well-written and reliable (ie. they do not often crash). There are lesser services that, alas, are sometimes required in the real world (and even the well-written ones may fail sometimes), and if you're running anything but a non-critical hobbyist machine, proper service supervision is not something that you can just write off as superfluous just because you don't like systemd. There *IS* a void in SysV init -- pretending it isn't there is just burying your head in the sand. There are many other things I do not like about systemd but I'm sick of the "everything is perfect" attitude; everything as it stands is *good enough* for me and I haven't found a replacement that is *good enough* in the same ways, but that doesn't mean SysV init is the end-all be-all init system, and it certainly doesn't mean that systemd has no merit whatsoever!
Quote:
Originally Posted by ReaperX7
Fadware has no place on UNIX and UNIX-Like systems on any level.
How can you possibly know whether or not something is just a FAD going in? You can make guesses, sure, but no one really knows until it is replaced. Slackware has deliberately excluded PAM for years but if you call PAM a FAD then you're delirious. It is used basically everywhere else and has been for a LONG time. pulseaudio has been around for a while now and with some projects removing support for raw ALSA output, I doubt pulse is destined to be a FAD either. Just because you don't like something doesn't mean you can dismiss it as being fleeting.
If sysvinit is good enough why are there so many alternatives.
... because people wanted to see if they could come up with something even better, obviously. The fact that it is so easy to make an alternative shows how good the sysvinit and core UNIX design is. Now, ask yourself how many alternatives to systemd are there? Is that because systemd is so perfect that no one can see anything to improve upon? Or for some other reason, such as the systemd project being designed to lock-out any alternatives in one big infrastructural land-grab?
systemd may be unavoidable at this stage. Doesn't mean we shouldn't resist to the bitter end!
Many years ago I turned my back on MS Windows and all it offers in pursuit of something better. Giving up GNOME or KDE will be trivial in comparison, and if linux nolonger provides that "something better", then I dare say I can adapt again.
is the likeability or otherwise of an individual developer a useful criterion for judging the usefulness of a piece of software?
Amongst us hairless apes, for good or for ill, conformity to some social expectations does matter. What little I've personally read from Poettering was a bit off-putting. If his judgment is bad for social dynamics, it does not necessarily follow that his judgment on technical issues will be bad but many will assume so. Get enough consensus that a person isn't worth dealing with and you get a fork or a project losing supporters. Some BSD forks are supposedly direct results of this. I've been involved in projects that died due to social issues.
That said, systemd has some design goals that are definitely interesting and probably worth pursuing. For my personal use, however, it's marginal utility at best. For others, it's demonstrably worthwhile. For the moment, I'm hoping that I can keep using a few specific KDE apps without it and XFCE as a whole won't be infected any time soon. Hrm, maybe it's time for me to start playing with Trinity again...
This paragraph is an excellent advertisement for systemd. Cron is not a service supervisor! On SysV init, if you need to restart a service when it crashes, you have a few options.
1) You can wrap the service in a script that relaunches it if it dies (using `wait $!` in bash or something more complicated in another language). This has the overhead of bash/ash/etc. and unless you're particularly clever it may require a different script for each service...yuck! You're also stuck playing around with PID files so you can properly stop the service without causing it to restart. I have done this myself but it isn't pretty and it certainly isn't ideal. It is also unlikely to work well with Slackware's existing init scripts so you may need to write a start/stop/wrapper script for every important service. Ew.
2) You can use cron to monitor a service at regular intervals and restart/take action if something is wrong. This is really ugly -- if the service is one that is expected to be up at all times, this solution is not usable (unless you want the service to be down for up to a minute, or longer if the check is performed less often). This also means that a check, using up resources, will have to run every minute -- if many checks are all running at the same time (or within seconds of each other anyway), this may cause a regular hiccup, which is unacceptable. If you decrease the frequency of these checks to mitigate that problem then there is a longer potential downtime.
3) Monitor it yourself and restart it when you notice it is down. For hobbyist servers this isn't really as bad as it sounds, but it certainly isn't an option for anything serious.
If sysvinit is good enough why are there so many alternatives.
If SysV init isn't good enough, why haven't any of those alternatives taken hold? Distributions didn't adopt systemd because it's so very much better than SysV init; they adopted it in the wake of it becoming a dependency of udev. Lennart had to weaponize software dependencies to get anyone (other then RedHat, his employer) to adopt it.
Quote:
If PAM is a SCAM why everybody uses PAM+LDAP. They could happily sync thousands user accounts between thousands workstations.
PAM makes centralized authentication easier with nontrivial (read: corporate) and heterogeneous infrastructure. It also introduces more complexity and thus failure modes and security risks. This is a trade-off that makes sense for some people, but not others. Slackware tends to take the route of greater simplicity and lower maintenance burden, and that's a reasonable decision. Deploying PAM and PAM-enabled packages on Slackware systems is still an option, and also a reasonable decision for some people.
Quote:
If someone is happy with his large home network of three computers why the world outside of his basement should stop moving.
LinuxQuestions.org is looking for people interested in writing
Editorials, Articles, Reviews, and more. If you'd like to contribute
content, let us know.