Story 2015-07-26 FGW5 Who's Afraid of Systemd?

Who's Afraid of Systemd?

by
in linux on (#FGW5)
Now that systemd is uneventfully running the latest releases of major distributions like Debian, Fedora, and Ubuntu, you might imagine that opposition to it is melting away -- but you'd be wrong. Instead, the rumors are as common as ever. Devuan, the anti-systemd fork of Debian, is still trudging towards a release while making the same arguments as ever. Devuan's home page asks: Have you tried to opt-out of the systemd change in Debian and stay with sysvinit? You will quickly notice that "Debian offers no choice." Yet a search quickly unearths instructions for making an install image without systemd and for removing systemd from your system.

Nor does the claim that systemd violates the Unix design principles stand up under scrutiny. Systemd is actually a general name for a series of related, similarly structured commands. From this perspective, systemd conforms to the principle of one program doing a single function in much the same way as the Linux kernel or a command line shell does. It is a suite of programs, not a single monolithic one. Systemd may not be ideal, but systems continue to boot and function the way they are supposed to.

In fact, not only are the most common anti-systemd arguments easily discounted, but they are surrounded by a vagueness that raises suspicions. Wild claims are made without any attempt at substantiation. The result is an air of secrecy and danger that, however appealing and reminiscent of freedom-fighting that it might be, does nothing to justify the anti-systemd rhetoric or make it plausible. Devuan's mailing list mostly shows the same dozen or so posters, and has raised only 7934 Euros. Supporters sound as though they are doing more fear-mongering than constructive effort.
Reply 60 comments

Might be a bad idea (Score: 4, Interesting)

by evilviper@pipedot.org on 2015-07-26 21:22 (#FGWD)

At the risk of reigniting the flames, I think |. should have at least one systemd story that isn't negative and conspiratorial. Besides, maybe the issue can be hashed-out in on-topic comments, instead of spilling-over into all other stories.

Re: Might be a bad idea (Score: 0)

by Anonymous Coward on 2015-08-03 07:27 (#G8HT)

Great, one. Where? A friend of mine who runs Linux as his home OS is now seriously annoyed. A recent upgrade included SystemD. Something borked the video card driver. The interupt for boot isn't working. Nothing could fix the issue. Problem diagnosis was painfil. Liken this to regressing back to Linux of 1995. He solved the problem by regressing back and now has Upstart working. A huge waste of his time. For what? How does SystemD help here? Screwing with the init so users can't diagnose problems? Stuffing up the init order causing fatal boot? This is an end user. Today he has removed the problem - Systemd. Tomorrow he is looking for a new OS or an easy way to get rid of it.

Re: Might be a bad idea (Score: 2, Insightful)

by billshooterofbul@pipedot.org on 2015-08-03 14:21 (#G9MP)

What do you mean by " the interrupt for boot" ? If its not getting to grub, or gummiboot that's not a systemd issue. Sounds like the update screwed up things. Rather than systemd. He's probably trying sysvinit style debugging techniques for systemd.

Re: Might be a bad idea (Score: 0)

by Anonymous Coward on 2015-08-03 09:50 (#G8WW)

At the risk of reigniting the flames I ask whether posts pro systemd could be less repetitive...

http://mobile.datamation.com/open-source/whos-afraid-of-systemd.html

Re: Might be a bad idea (Score: 0)

by Anonymous Coward on 2015-08-03 09:59 (#G8X8)

uneventfully?? Just because people are not complaining on boards lists and usenet does not mean that this is uneventful.

Re: Might be a bad idea (Score: 0)

by Anonymous Coward on 2015-08-03 10:12 (#G8YK)

"most users are unlikely to make a fuss so long as their systems continue to boot and function the way they are supposed to." If systemD dies, then this is likely to be why.

from the sidelines (Score: 3, Insightful)

by pete@pipedot.org on 2015-07-26 22:14 (#FGZ2)

with only scant knowledge of the implementation of systemd, i'll continue to be wary; I'll say, in large due to Linus' response to Kay Sievers requesting the kernel be patched to fix his own bugs.

I don't hold Linus in too high regard because of his often unnecessary crass, but he usually makes a great point. I've read a couple examples of cockyness and ego, and for me that raises red flags for potential vulnerability or simply poor coding - not much is more dangerous than a programmer who believes they don't make mistakes. I'm open to counter-examples tho, i'm not anti-systemd because of some movement, perhaps its simply this

Re: from the sidelines (Score: 1)

by billshooterofbul@pipedot.org on 2015-08-03 17:06 (#GA5B)

Getting yelled at for breaking a userspace app and or kernel debugging features is kind of a mark of honor. Alan Cox got reamed for breaking emacs a while back. Not wholly undeserved, but its kind of Linus' role to protect userspace from kernel devs and kernel devs from userspace. Its easy to get tunnel vision and only care about a particular deep logical problem and find a way to fix it, but in such a way as it breaks compatibility. Of course when egos are involved, sometimes people don't put the finest foot forward, but eventually cooler heads prevail and solutions are found. In Kay's case, it was a fight over boot params namespace. I kind of agreed with him on that, but the change they did make really killed kernel debugging, and so while he might have been technically correct, it was the wrong way to do that due to the impact.

Getting on like a house on fire (Score: 2, Interesting)

by Anonymous Coward on 2015-07-27 00:17 (#FH5W)

Perhaps the question should be: In the face of continued great ongoing resistance against a new init system why is it being shoved down everyone's throats? I have looked at systemd. I want no part of it. At all. It looks and feels like something from microsoft.

Re: Getting on like a house on fire (Score: 2, Insightful)

by axsdenied@pipedot.org on 2015-07-27 07:17 (#FHWQ)

As mentioned in the article, what ongoing resistance is actually there? I am in no way expert on systemd but the whole thing may be being blown out of proportion by few people who don't like systems or Sievers. Or who are resisting the change.And I keep wondering why would the more knowledgeable people (who make decisions at Debian etc) agree to go with it if it was that bad.Edit: Accidentally posted as reply to wrong post. Not sure how to fix...

Re: Getting on like a house on fire (Score: -1, Redundant)

by Anonymous Coward on 2015-07-27 07:59 (#FHZX)

You make a good point. You are no expert. Okay. Your care factor may be finite compared to people who administrate production machines You could grab a copy of VM software, download some BSD and Linux ISOs to have a play. Screw the system up. Break it. Fix it. Install software. Get it working. Imagine doing this across 50 boxes. Now imagine 500. 5000. Make changes in a sysv init sequence. Make changes in a systemd system init. Learn what people are talking about. It's fun!

Re: Getting on like a house on fire (Score: 2, Insightful)

by axsdenied@pipedot.org on 2015-07-27 15:30 (#FK5H)

Perhaps next time try addressing at least one of the real points I made.

Re: Getting on like a house on fire (Score: 1, Interesting)

by Anonymous Coward on 2015-08-03 09:14 (#G8SE)

Debian is indeed what raised a red flag on systemd for me, because the debian way should have been to bring up a dialog during the installation with sysvinit vs systemd choice. And both would have ended up in a system that is mostly working well. The inability of achieving that means either: political pressure was applied, see also the spotify request that turned out as a joke on debian, or that systemd is too invasive from an architectural POV. I gave my best shot at liking it anyway but i have yet to see any feature of my intetest that a system equipped with runit and traditional tools cannot do better. To me it seems a trick to train new devs to update systems just for the sake of it. Commercial IT philosophy done with open source.

Re: Getting on like a house on fire (Score: 1)

by billshooterofbul@pipedot.org on 2015-08-04 20:14 (#GE2P)

Everything was done in the open, look at the mailing lists to see what transpired. It was all in the open. Systemd is a better solution. Making it init by default is a sane choice. Other init systems still work, but giving devs a single one to ensure works makes their lives easier.

Systemd isn't invasive, its fundamental. Its the cgroups manager which provides an api for cgroups for other things to use. That just makes sense to have init be the cgroups manager.

Re: Getting on like a house on fire (Score: 2, Interesting)

by dnied@pipedot.org on 2015-07-27 19:48 (#FKZ0)

It looks and feels like something from microsoft.
Or Apple. I'm forced to keep a Mac. It has "launchd" and I freaking hate it with a passion. Especially because it subsumed cron. Now, in order to schedule jobs, you have to write lines upon lines of fucking XML. No way.

Re: Getting on like a house on fire (Score: 1)

by billshooterofbul@pipedot.org on 2015-07-28 14:21 (#FPHY)

True statement.

launchd inspired upstart.
Pottering wanted to modify upstart, but the copyright assignment wasn't cool with him or his company. So, systemd is really pottering's rewrite/refactoring of upstart.

So yes, its very much Mac like in that regard. But on the plus side, the config files are pretty straight forward ini files ( ok it has a little MS inspiration too).

Re: Getting on like a house on fire (Score: 1)

by lmariachi@pipedot.org on 2015-07-29 10:29 (#FSGA)

You don't have to do it that way at all. You call whatever script you want from your LaunchAgent and tell it when to fire, and you can do it via a graphical XML editor if that's such a hurdle.

Or you can put your scripts in /etc/periodic/(daily|weekly|monthly) if you want anacron functionality.

Or you can just keep on using cron! It is called by launchd, but that shouldn't make a difference. Point is, you're not forced to delve into launchd or write "lines upon lines" of XML to schedule jobs.

Reading those paragraphs (Score: 1, Insightful)

by Anonymous Coward on 2015-07-27 02:15 (#FHBJ)

Makes me think there is something wrong with systemd. All the assertions that you make to try to defend it seem manipulative. It's like whoever wrote it actually knows there is something wrong with it.

Re: Reading those paragraphs (Score: 0)

by Anonymous Coward on 2015-07-27 13:57 (#FJXH)

Where there is smoke...

Re: Reading those paragraphs (Score: 3, Interesting)

by Anonymous Coward on 2015-07-27 20:49 (#FM4C)

It's not clear that there is actually much wrong with systemd except the binary logging. But it's not clear that there isn't. And there were lots of horror stories about unrecoverable errors around a year ago. Perhaps they've addressed it since then.

OTOH, init was working perfectly fine in my use case, and I'd really rather not have switched. And systemd seems intent on swallowing so much stuff that nothing will work on both an init based system and a systemd based system. They aren't there yet, but...

So I don't trust the developers. And I'm upset at being coerced into using it. I *am* using it, but I'd really rather not be doing so. Systemd seems designed for administrators who administer a large number of servers/workstations/pcs. That's not me, and I'm unhappy at being coerced into using it. If Devuan ever gets its act together, I may well switch.

Re: Reading those paragraphs (Score: 0)

by Anonymous Coward on 2015-07-28 05:32 (#FN58)

Binary logging is a myth. You can still read the logs with any test editor. You can also use jourald for some nice features, but it is not required; the logs are not binary.

Re: Reading those paragraphs (Score: 1, Interesting)

by Anonymous Coward on 2015-07-28 15:32 (#FPT5)

Just out of curiosity, what is this talk about systemd being targetted at maintainers with lots of machines? Does it have some kind of remote management interface? I keep hearing about this but the systemd webpage isn't very helpful. Any quick comments?

Re: Reading those paragraphs (Score: 1)

by evilviper@pipedot.org on 2015-07-28 17:49 (#FQAG)

It can be configured to automatically restart services that exit, provides features that support configuration management engines like Puppet, automatically handles service dependencies, and more.

You can also see "A short list of other features:"
* http://0pointer.de/blog/projects/systemd.html

Re: Reading those paragraphs (Score: -1, Troll)

by Anonymous Coward on 2015-07-28 20:09 (#FQPX)

It looks like Puppet can handle anything you throw at it, if you throw it well, not just systemd. Are you talking about the d-bus interface?

I read most of that post, but I still don't understand who this is intended for. Fast booting doesn't make sense for servers. They are not booted frequently, it's fine if a server boots in 10 minutes instead of 2. What matters is that it boots consistent with its intended configuration. That post talks a lot about how starting services in parallel speeds up things, and most importantly lazy activation. Lazy activation is unacceptable, at least in our setup. We'd like to know of a failure (most often misconfiguation) as soon as possible at boot time, not when we actually need the service.

It doesn't make sense for workstations either. On workstations, there is often NO service running. We have the servers for that. Booting a workstation fast is a piece of cake, really. Mount your drives, get a network address, start up a couple daemons like the logger and the console and you're done. You don't need a complicated, hard to configure program to do that. Unless you redefine X11 display to be a service, the mouse driver to be a service etc. It doesn't make sense. Help me with this.

It very much looks like Lennart is bored with fucking up the audio system and is trying to fail at something new.

Re: Reading those paragraphs (Score: 0)

by Anonymous Coward on 2015-07-29 14:17 (#FT5C)

That is a good read. In no way does it help to make systemd more palatable. If anything reading this makes it worse.
Far worse.
There are some good idea there. Pity about the implementation.

Re: Reading those paragraphs (Score: 1)

by billshooterofbul@pipedot.org on 2015-08-03 14:24 (#G9N7)

When you have a large number of machines, the odds of you hitting the flaws inherent to sysv init and upstart become more frequent. systemd has a large number of "Proper" solutions to things that sysv and upstart work for 98% of the time. It also has a number of things that make administering systems easier, faster ( if you take the time to learn them instead of sticking your head in the sand and complaining about having to learn new things).

Re: Reading those paragraphs (Score: 1, Interesting)

by Anonymous Coward on 2015-08-03 09:28 (#G8TV)

The other wrong with systemd can be easily seen searching for his hashtag on twitter. I was struck by the attitude of Poettering that had been asked by google not to use their time servers in the dev systemd branch for both resource utilization and accuracy reasons. The guy first replied that he had no alternative like pool.ntp.org because he is not a vendor and systemd is not a product. OK, we know that, but that should not interfere with google request., nor should be a problem for the guy on RH payroll that is redesigning core part of linux based systems. Look at the thread yourself, it is fascinating.

Same old (Score: 0)

by Anonymous Coward on 2015-07-27 08:02 (#FJ0C)

Systemd may not be ideal, but systems continue to boot and function the way they are supposed to.
Thank you for that, I rarely laugh at anything on the Internet.

This is exactly what the problem is, SysV-init already worked perfectly before systemd. systemd doesn't bring anything to the table except for complexity and more dependence on the Redhat and Gnome group. That's why we oppose.

It only makes the job of OS distributors easier, that's why it's accepted all around. Debian guys no longer need to write any init scripts, they can just copy them from Redhat. That's why. Users opposing systemd are all bitten by it in some way or another. They criticise, and try to expose the flaws of systemd, but users' opinions don't matter in the free software world. Since no money changes hands, the developers are free to do whatever they like. See how GNOME went down.

Yeah, another funny thing. Imagine you go to a car dealer and the salesman pitches you some car. You tell him you don't like some aspect of the car and the salesman tells you "Let's see you do any better!". This is what it is.

People aren't doing systemd replacements because:

1- It's a bad idea. The goals of the project are broken, why should anyone try to imitate shit when they can already get shit ready made.

2- Surprise surprise, SystemV-init still works. If you run an older udevd (or eudev, I haven't checked it out) without systemd support, there is no impact on anything really. On my simple desktop machine at home, I have exactly one init script for the system. That's my replacement for systemd.

3- Nobody cares enough about this to allocate significant resources to it. You see, in the free software world, whoever commits his resources wins. It doesn't matter whether they are doing a good job or not. It reminds me of the demise of the Bluetooth support under Linux.

Shitty stuff is always like this, once its starts going around, it generates lots of talk. We could do a whole big discussion about how systemd fails but what's the point? Systemd proponents will still go "la la la I don't hear you, everybody is using it".

Re: Same old (Score: 1)

by Anonymous Coward on 2015-07-27 19:22 (#FKX2)

People also forget the Redhat's target is enterprise and enterprise is incredibly slow. My relatively small group isn't scheduled to finish testing a version with systemd until the end of the year.

There will certainly be some huge corporations that are upset about the switching costs, and how they are handled will be what decides systemd's future.

Re: Same old (Score: 1)

by Anonymous Coward on 2015-07-28 14:27 (#FPJ0)

We are switching to RH soon. Regardless of init system. If SystemD causes problems there will be Hell to pay. By us. Against us. Redhat could cause hundreds of people to avoid them if this borks badly. People carry grievences for a long time. We shall see what happens. I step forward with eyes and mind open.

Ubuntu (Score: 1)

by zenbi@pipedot.org on 2015-07-27 08:09 (#FJ12)

Probably better for Ubuntu than Upstart. They never really finished moving from SysV Init scripts to Upstart Init scripts. At least now they have one less one-off NIH project. Now if the Wayland people can convince them to give up on Mir...

My quibble with systemd is with the binary log files. There may be some workarounds to get normal text logs out of the journal, but I'd rather not have to do such things by default.

Re: Ubuntu (Score: 1)

by billshooterofbul@pipedot.org on 2015-07-28 14:24 (#FPHZ)

I really wonder what's behind the binary log complaints. If the format is open and the file gets corrupted, its still easily to figure out the content. Its not encrypted

Re: Ubuntu (Score: 2, Insightful)

by Anonymous Coward on 2015-07-29 08:07 (#FS4P)

Corruption is rarely relevant when it comes to logs. Logs in important systems are frequently archived, and as any backup they are stored in a reliable manner. The single most important thing about logs is how they are USED. If you just let the logs sit on your HDD, then there's no point in collecting them (increase your logging level to ERROR and be done with it). I, on the other hand, use logs to audit systems continuously, and if some software is introduced that makes it more difficult to analyze the logs then it is a serious problem for me.

The problem I have with journald is that it makes handling the logs more complicated, without providing any benefit compared with alternative solutions. The alleged benefits (e.g. binary logs) are already implemented in all other systems, and those implementations are more thought-through than those of the SystemD-team, as each development team can focus on polishing a single tool rather than a beowulf of programs (this is the Unix-philosophy in a nutshell). Storing logs as binary files, for example, is either done using logrotate (for archiving in compressed files) or by forwarding selected parts of the syslog stream to a database (for future analysis).

Therefore, the only way to make useful log analysis with JournalD is to configure it to forward everything to regular syslog, and add your regular analysis tools to the syslog stream. This means that in a system with both log analysis and SystemD, journald is an extra step in your pipeline which doesn't contribute with any functionality. In addition, this means that all logs (which are generated as text) must be converted into binary only to be re-converted back to text before being forwarded to analysis tools.

Now you're probably saying that tools can read the binary files straight-up in order to make things more efficient. This is true, but the problem is that there are programs which log to their own logfiles, rather than passing their logs via the syslog functionality. This means that the log analysis solution must still be capable of handling text files. This means that you must have support for both the binary format and the text format, rather than (as with other solutions) just handling text files. This (once again) increases complexity without adding any value, which is bad.

Tl;dr: Fuck Journald.

Re: Ubuntu (Score: 1)

by billshooterofbul@pipedot.org on 2015-08-04 17:02 (#GDGG)

That is a unique argument that I have not heard. I don't find it to be very persuasive. Journald is better than syslog and other alternatives in many ways. But it is a change, and as such there will always be people who cry "change for no reason!!!" because they don't see or understand the benefit from the change and have to make changes to their existing work flow. There are reasons, and they are documented. The only relevant argument against has been existing tooling needing syslog.

What is this? Slashdot? (Score: 0)

by Anonymous Coward on 2015-07-27 12:20 (#FJMW)

Look, it's great that you've got an opinion, but if you're going to post it as news, have the decency to say "EDITORIAL" at the top of it.

This site is getting a well deserved reputation for posting high quality submissions, and if the replies are few, they're generally thoughtful. It wouldn't take many posts like this to destroy that entirely.

On the other hand, if you're going to start regularly trolling for eyeballs, just say and I'll go elsewhere. I had enough of that on /.

Re: What is this? Slashdot? (Score: 0)

by Anonymous Coward on 2015-07-27 13:44 (#FJW9)

Reviving the discussion for a good old PIR is always welcome. Has systemd won? Are users moving away to nonsystemd platforms? Will this impact the enterprise linux users? Does anyone care?

Frankly, I am more concerned about the TPP right now.

Re: What is this? Slashdot? (Score: 1, Informative)

by evilviper@pipedot.org on 2015-07-28 04:33 (#FN11)

No, if this was /. this would be yet another ranting anti-systemd story... they love to fan the flames no matter what the facts might be. This is also not a |. editorial at all, either, click through to the source for yourself. Note all those pesy facts and figures...

Re: What is this? Slashdot? (Score: 0)

by Anonymous Coward on 2015-07-30 11:01 (#FX33)

No, if this was /. this would be yet another ranting anti-systemd story...
Yes, yes, yes. the shortcomings of Slashdot are well known. I visit this site in the hopes that it might not share them.

"If you fight against monsters you should take care that in the process you do not become a monster yourself".
click through to the source for yourself. Note all those pesy facts and figures...
Errm, what facts, exactly? All due respect to Bruce Bryfield, but he posted a blog entry presenting an unsupported opinion piece. The only figures present are where he reports the total donations received by Devuan to date. The only facts I can see apart from that are that Lennart Poettering is involved and a lot of people don't like him very much.

He doesn't even try and summarise the arguments pro-systemd, let alone the ones against. It's hardly a ringing vindication for either side of the debate. I don't think it was ever intended to be.
This is also not a |. editorial at all, either
You posted a highly partisan opinion, presented it as news and backed it up with one solitary link to an equally partisan blog post. If you'd said "Bruce Bryfield has posted an interesting article advocating acceptance of systemd. I think he makes some good points, what do you think?" then I wouldn't have an objection. Instead you posted one saying "I'm right, anyone who disagrees with me is wrong, and by the way this dude (who I can't even be bothered to name) agrees with me, so that settles it".

I'm sorry, but regardless of whether your opinion is right or wrong, you're still editorializing. And I'm not saying you shouldn't be able to do that. I just ask that you draw a clear line between your opinions and what you're reporting as news. I hadn't thought it was a lot to ask.

Re: What is this? Slashdot? (Score: 1)

by evilviper@pipedot.org on 2015-07-30 19:54 (#FYTQ)

Actually, it's the OTHER systemd stories that hit |. that were slanted opinion pieces. Quite a few of them, too. I don't recall you complaining about any one of those. This doesn't qualify, because the claims made are factually supportable and settled, even when just correcting misconceptions and inaccurate arguments previously used. That you still can't see any of that says worlds about your own bias, and I doubt anything else I can say to you will help.

Re: What is this? Slashdot? (Score: 0)

by Anonymous Coward on 2015-07-31 07:28 (#G06B)

My complaint isn't about systemd. I've made that clear from the start. It's a question of journalistic integrity.

If you can't see that, then you're right - nothing you can say is going to be very helpful, and I'm wasting my time and yours both.

Good luck with the site.

Re: What is this? Slashdot? (Score: 1)

by billshooterofbul@pipedot.org on 2015-08-03 17:20 (#GA6J)

Sadly, Journalistic integrity seems to be of the exclusive domain of the NYT and possibly the Economist. Everything else is shades of grey. I typically don't expect much from any other source. Good journalistic writing is a skill that must be learned and honed. Its not found on any random websites that happen to post news articles.

Re: What is this? Slashdot? (Score: 0)

by Anonymous Coward on 2015-08-02 18:31 (#G77Y)

Errm, what facts, exactly?
Exactly. Whenever some systemd proponent tries to market their shiny, they never bother with the facts. Look at the posts in this story. Anti-systemd guys talk about some serious problems with systemd and the proponents simply go, "Bah, you don't know better than Debian people". No addressing of the facts, nothing to back it up. Nothing.

This has been typical response in all the stories about this old issue. This is because it's just marketing, there is nothing to support systemd. Maybe in the future they can implement a halfway decent system but for now, it sucks. They know it and are in denial.

Re: What is this? Slashdot? (Score: 3, Interesting)

by fnj@pipedot.org on 2015-07-28 04:34 (#FN1T)

Agreed the submission is very slanted, but many systemd submissions have been slanted. I'm not slamming it hard or anything. The "who's afraid" part is a loaded term, but that may not have been the intention. A non native english speaker for example would not necessarily be expected to be very proficient with the nuances of various terms.

OK, here's my input. I'm not "afraid" of it but I still resent the steamroller, and I have grave misgivings about rolling up too much stuff in it. But I use it every day - I'm running Arch on my desktop.

I have gone further out of my way than I already was to get experience and familiarity with FreeBSD, spurred in part by the weakness in linux design directions revealed by the systemd steamroller.

Finally, yes, I am aware that "steamroller" is a loaded term also :-)

Re: What is this? Slashdot? (Score: 0)

by Anonymous Coward on 2015-07-30 12:08 (#FXAH)

I suppose my point is that the post isn't about Bruce Bryfield's article. It's primarily evilviper's opinion, posted as news with Bryfield's piece linked to in passing.

I'll admit to being in the anit-systemd camp, but that's not what annoyed me here. It's posting opinion and advocacy as news. I don't mind anyone having an opinion, but a responsible news outlet needs to distinguish between reporting things that have happened and the editor sounding off about whatever has them steamed up at any given moment.

And the fact that most of the systemd debate has been very low grade doesn't really excuse |. from joining the race to the bottom. Or maybe it does: it's a question of what sort of site Pipedot wants to be.

Personal choice... (Score: 2, Interesting)

by timex@pipedot.org on 2015-07-29 19:23 (#FV3Z)

For ME, I don't like the way that systemd is marketed as a "sysVinit replacement", but systemd has hooks into all sorts of things, including (but certainly not limited to) system logging and GUI applications.

If sysVinit is going to be replaced, let it be replaced by an actual init system, not something that has more in common with mold, getting its tendrils into everything imaginable.

That is, admittedly, a personal opinion. I'm currently running Devuan on a netbook and I'm quite happy with it. Would I *use* another distro? Sure, if I had to. If I have a choice though...

Re: Personal choice... (Score: 1)

by evilviper@pipedot.org on 2015-07-29 20:08 (#FV81)

While a reasonable point, you certainly can't claim Linux was sleek and simple and clean before systemd. For many years there has been a huge mess with ConsoleKit, NetworkManager, avahi, dbus, dcop, hal, pam.d, udev, devfsd, sysfs, +proc, devtmpfs, kudzu, zeroconf, and much more crud. It seems to me that when USB came along and make extreme plug-and-play user-facing, Linux distros just kept throwing in everything but the kitchen sink to make each scenario work right for ignorant desktop users and GUI management tools (and a nightmare to configure and debug for administrators).

With systemd starting to take over bits of some of those, I'm hoping for a bright future where there's one little simple place where such device configuration can be done (and those being the same across distros, and also not changing between each major release), and not several that do different but partially-overlapping bits of the puzzle, as we have to tolerate now. Maybe systemd is not actually be the best answer to that problem, certainly the BSDs manage to work fine without the Linux insanity I listed above, but at least it is finally some progress towards unification and simplification of the complete mess.

Re: Personal choice... (Score: 1)

by timex@pipedot.org on 2015-07-30 16:43 (#FY7C)

you certainly can't claim Linux was sleek and simple and clean before systemd. For many years there has been a huge mess with ConsoleKit, NetworkManager, avahi, dbus, dcop, hal, pam.d, udev, devfsd, sysfs, +proc, devtmpfs, kudzu, zeroconf, and much more crud. It seems to me that when USB came along and make extreme plug-and-play user-facing, Linux distros just kept throwing in everything but the kitchen sink to make each scenario work right for ignorant desktop users and GUI management tools (and a nightmare to configure and debug for administrators).
I don't recall claiming that Linux was "sleek and simple and clean" before systemd. I know things can get messy and complicated (I'm a *nix system administrator by trade). I'm simply stating that in my opinion, systemd is not necessarily the solution that should be sought.
Maybe systemd is not actually be the best answer to that problem, certainly the BSDs manage to work fine without the Linux insanity I listed above,
Well... The BSDs manage, but they have their own problems. FreeBSD has (from my view) a serious problem with software availability. If you wanted to add something and it wasn't in the official repository or in Ports, you have to get the source, else you are stuck. If you're lucky, you might be able to get the Linux binaries, and there are libraries to allow you to run most Linux-compiled programs on FreeBSD. I don't have more than a cursory experience with NetBSD or OpenBSD, but I would imagine the issues there are similar.
...but at least it is finally some progress towards unification and simplification of the complete mess.
I might point out that many issues related to the hodge-podge nature of drivers and the like are due to the major distributions' approach to inclusion. Instead of Red Hat/Fedora making an official repository available with vendor drivers for certain hardware (video cards, in particular) or codecs (*cough*MP3*cough*), they have chosen to refuse to make them available at all. This leads people to find unofficial repositories, some of which can be trusted, some which no sane admin could trust.

Either way, if systemd is supposed to be a tool that replaces all sorts of things, I think it's a Bad Ideatm. Many can easily make a case that sysVinit is broken and needs to be replaced. Fine. There are several options available. I'm largely annoyed that systemd has forced its way into almost every distribution to the point where even if one does choose another init system, systemd is still active because it's required for other things.

"Unification"? Isn't that what Microsoft does?

security. (Score: 1)

by gravis@pipedot.org on 2015-07-29 22:26 (#FVJS)

Systemd may not be ideal, but most users are unlikely to make a fuss so long as their systems continue to boot and function the way they are supposed to.
my concern has been and is still about the security of systemd. it's a huge amount of unproven code and it's running as root at all times. when someone finds a remote exploit, it's going to be worse than heartbleed because it will be a backdoor into a large number of servers. will people care then?

Re: security. (Score: 0)

by Anonymous Coward on 2015-07-30 14:23 (#FXR5)

We are screwed. Plain and simple. On the bright side, I have learnt how to strip it out of Debian manually. We don't need a Debian fork. We just need a tool to remove and disable systemd.

I'm free (Score: 4, Interesting)

by zafiro17@pipedot.org on 2015-07-30 18:14 (#FYGF)

Switched to PC-BSD from openSUSE earlier this summer and have no regrets. It wasn't a blind install - I've used it before and like it. But PC-BSD gets you a working FreeBSD install with minimum fuss, and unless you have the bad luck of hardware that FreeBSD doesn't like (laptops and wifi cards can be a hassle), you've got a sweet *nix system that does everything your Linux distro did, and no systemd. A lot of other conservative choices, too.

I'm not writing to gloat. Just pointing out: if you don't like systemd, switch your operating system. Pretty sure Slackware also avoids systemd. Vote with your feet, people.

Re: I'm free (Score: -1, Troll)

by Anonymous Coward on 2015-08-03 19:51 (#GAMH)

I did something similar a couple of months ago, but to Windows instead of BSD. What a refreshing change! My sound card finally works right, it doesn't cancel line-out forever when I plug in a headset. Bluetooth works, and most importantly, the video card works with all the acceleration it has. I'm finally actually using the hardware I paid for. Besides, there are nifty programs I could never use with Linux.

I do miss a couple of things from my unixy setup, but life is too short to try to make my operating system at home perfect.

I still boot into Linux or run it thru a VM occasionally, but that's just for using the tools I made over the years. How I thought I'd be a UNIX guy forever..

Re: I'm free (Score: 0)

by Anonymous Coward on 2015-08-04 15:47 (#GD96)

wow what version of windows you are using? last time I used it, it was a win8 luckily configured as a traditional desktop (else I'd be probably still be there figuring stuff out), it was to restart an intranet proxmox instance who apparently needed console access because the disk was encrypted and asked for the password at boot.

Fire up explorer. Where the f. is the address bar? oh at the bottom, OBVIOUSLY. I type the address, nothing happens. Not even an error. I check the address, it is correct.

I see a chrome icon. Let's try that.

Chrome doesn't like the certificate. Good chrome, sit, it's a f.in intranet, sit. Clicking through a couple dialogs sets things straight.

Enter proxmox web interface, click for the console... no java plugin available. Download? activate? well, those kind of plugins are not liked anymore, so they will be installable till version XYZ and later no way....

Chrome, damn you to hell, it's a f.in intranet!

Raised hand: "client, I need to install 200mb of java and 25 of firefox to click on a fucking link and have a fucking terminal window, or i have to spend between 2 and 120 mins on proxmox docs and setting up alternative access from win to the VM. Else I'd boot a live linux from usb, install whatever is needed if it's needed in RAM, and GTFO in 10 mins without touching one bit of your workstation HD. WHAT WILL IT BE?"

The client was already having convulsions at the thought of the java plugin popups returning in his life, so it was an easy sell.

Re: I'm free (Score: 1)

by hyper@pipedot.org on 2015-08-04 23:02 (#GEG7)

What is your choice of live cd boot distro for this scenario? Ubuntu?

Re: I'm free (Score: 0)

by Anonymous Coward on 2015-08-09 22:40 (#GXM5)

In that occasion I used a recent kubuntu, probably w/systemd. It sucked (does not apply setxkbmap, no error)but did the job.

Re: I'm free (Score: 0)

by Anonymous Coward on 2015-08-10 02:39 (#GXZ1)

Thanks for the info. I keep several linux and bsd live cds around for this purpose. Moving slowly to live-usbs.

Devuan needs large iso files (Score: 2, Interesting)

by Anonymous Coward on 2015-08-07 03:39 (#GP6M)

Devuan is starting to look like it's repeating Debian's mistakes. 1)It is trying to support lots of architectures up front. Why not stick with x64 for now, handle the other architectures later?
2)lots of file formats, but the only ISOs are for netboot. That's OK for people with lots of bandwidth. If they built a large (600MB or bigger) ISO I could download it while I sleep (which is the polite thing to do since my bandwidth is limited) and use it for multiple installs. Especially if the ISO included language packs. Installing Debian is a real PITA and takes forever because it always downloads those language packs. That's dumb. 3)take the time to make a liveCD. It isn't that difficult. 4)create an ISO image that can go on a thumbdrive with an NTFS filesystem, using Unetbootin. The 'dd' command is destructive. It's another PITA to reformat the thumbdrive to make it usable after the install. This is part of 'plays well with others'.

Systemd is trash. (Score: 0)

by Anonymous Coward on 2015-09-03 14:32 (#KAW5)

Systemd is trash.

Everything good in linux is taken away.

Re: Systemd is trash. (Score: 0)

by Anonymous Coward on 2015-09-05 12:53 (#KH5A)

Yeees but some more information about this opinion would be appreciated

<3 gentoo (Score: 0)

by Anonymous Coward on 2016-04-02 03:02 (#192PH)

Well that is why I love gentoo because I have the choice to install systemd, but fuck systemd i'll take openrc over systemd any fucking day of the week and I can guarantee you that my system will boot faster than any fucking box running systemd.