Linux VS open source UNIX

https://www.adminbyaccident.com/politics/linux-vs-open-source-unix/

by rodrigo975 • 7 years ago

47 56 7 years ago

elagost2• 7 years ago

This weekend I started digging into OpenBSD. I'd never used a BSD - been a Linux user for the past decade. While the table linked might show OpenBSD as the "least complete", what seems to be the case is that they pick one tool to fit a use case and then use that one. There aren't 400 different shells installed by default or 5 different process monitoring/inspection programs, because they have the one they like and use/improve upon it.

OpenBSD is a fascinating operating system. I'm only a few chapters into Absolute OpenBSD (https://en.wikipedia.org/wiki/Absolute_OpenBSD) by Michael W. Lucas, but from a Linux user/admin perspective it is enthralling. I can't wait to learn more about it.

Enginerrrd2• 7 years ago

The systemd debacle is what led me down a rabbit hole toward the BSD's. Don't get me wrong, systemd actually adds a lot of needed functionality, but I really feel like it is a major philosophical branch away from the conventional Unix thinking of 'make a tool that does one thing really well'.

That mindset has such a strength in that such tools form reliable building blocks for other tools and system components in a way that is really approachable for amateur/volunteer developers. It takes no time to parse through the man page of a simple tool. And only a bit longer to parse the code. That makes it easy to build improvements, or build on top of it.

But trying to wrap your head around systemd in even a few hours isn't reasonable. And binary log files? Wtf. It also seems like relying on certain aspects of it's behavior is a good way to end up breaking your system.

BSD still has that old school Unix mindset that made Linux great in the first place. Especially up against Microsoft's products.

The proof in the pudding for me is to look at the running processes of a default system between BSD and Linux with systemd. Yikes.

dreich2• 7 years ago

The systemd debacle is what led me down a rabbit hole toward the BSD's.

Like most, you could have simply moved to a distribution without systemd. Unlike Linux, the BSDs offer less choice when it comes to both init, service supervision and service management options. It is also evidenced in the comparison table in the OP. If anything the debacle resulted in creating more awareness and leaving people open to more choice.

That said, adopting one monolith instead of another sounds more like a knee jerk reaction than anything else.

The proof in the pudding for me is to look at the running processes of a default system between BSD and Linux with systemd. Yikes.

This only scratches the surface. Daemons running under systemd have their PIDs supervised instead of using unreliable hacks like pidfiles. This is one of the reasons why systemd is preferred in large deployments and professional environments.

ti_ranger4• 7 years ago

The systemd debacle is what led me down a rabbit hole toward the BSD's.

As far as I can tell (not actually using Debian or Ubuntu myself), there was only a debacle in Debian and Ubuntu, because they took too long to decide, and ended up having to rush their migration.

All the other distros, that had better-planned migrations (Arch, Fedora, SUSE, etc. etc. etc.) had basically no issues migrating to systemd, and have benefited immensely from being able to ditch the maintenance overhead of diverging init systems and init scripts, all re-implementing the same new features with slightly different implementations due to previous slightly differing implementation details.

People will currently call out OpenRC as the one non-systemd approach that works (e.g. if you acknowledge that SysvInit no longer actually works because no-one uses if for all system initialisation, instead leting it delegate lots of control to other bits and pieces), but before systemd won, there were: * SysVinit + RH-derived initscripts * Sysvinit + SUSE-style initscripts * Upstart * OpenRC * prcsys - parallel rc system (Mandrake and derivatives)

All of these had similarities, and small differences, e.g. they all tried to support dependency information in init script headers.

As a sysadmin, it was frustrating to: * Have to read every init script to see where it read its defaults from, was it /etc/sysconfig/pkgname or /etc/default/daemonname . Did it support setting limits from the defaults file, or would you have to hack the script to support your bigger-than-the-maintainer-imagined environment. * Debug where the application was logging to, if it didn't setup logging correctly by default.

As a maintainer of a package used on more than one distro, it was difficult to: * Test your init script to ensure it worked on RH-based distros and SUSE and Debian and Ubuntu. While Upstart said it supports init scripts, it fails in some random way that doesn't occur on RH and SUSE. * Setup logging correctly, as not all distros provided a helper script to setup the right logging daemon (syslogd, vs rsyslog vs syslog-ng etc.)

We won't talk about the more modern features such as cgroups etc., where only about 2 init systems support the feature adequately, and no distro or packager actually managed to support such advanced features on more than one init system, and certainly not more than one non-systemd system.

systemd mainly stays out of my way, but makes it easy to troubleshoot and customise daemon behaviour in a consistent fashion.

But trying to wrap your head around systemd in even a few hours isn't reasonable.

You don't need to wrap your head around systemd to do things with it that you would really struggle to do with any other init system. But, you've probably been too busy fighting it, or trying to prove why it's worse, rather than trying to do real things with it, or you would be able to point to valid bug reports.

The proof in the pudding for me is to look at the running processes of a default system between BSD and Linux with systemd. Yikes.

You're comparing apples and oranges here, you should rather be comparing two release of the same distro, before and after systemd became the default, and you would have seen very little difference.

moviuro• 7 years ago

BTW, there is currently (until Nov. 18th 2019) a sale of BSD & Linux ebooks at humblebundle[0] Absolute OpenBSD and Absolute FreeBSD are in that bundle.

[0] https://humblebundle.com/books/linux-bsd-bookshelf-2019-book...

codetrotter2• 7 years ago

Humble Bundle used to make it possible to select which charity you wanted to give to, and I used to select the FreeBSD Foundation and give them the whole amount.

But I can't figure out how to do custom selection of charity any longer. Does anyone know?

elagost2• 7 years ago

That is what prompted me, actually! My coworker alerted me to this sale and I jumped on it. Something about Absolute OpenBSD spoke to me, and I had to pick it up first... maybe it's the "Practical Paranoid" part.

drivers99• 7 years ago

I see it's available in Safari Books (for those who have a subscription.) Wikipedia says, "the second edition was released in April 2013." Looks like that's just before version 5.3 came out, and there have been 13 releases since then. However, there are mostly 5 star reviews on Amazon, including ones from 2019, so it sounds like it's still worth reading.

unwind• 7 years ago

Meta/mods: please shorten the title, to remove the needless duplication of the site's name. Thanks.

jonfw• 7 years ago

This guy has a real knack for complaining about an issue, and than admitting it was actually fixed a while ago.

Complains about a certain guy's benchmarks being affected by using different compilers, but of course that guy switched to using the same compilers already

Than he does the same thing with epoll- complaining about something being broken, but then admits it was fixed in 2016

I wasn't really able to pull anything material out of this article. I guess this is why people like benchmarks- it's hard to write an opinion piece about performance.

dvfjsdhgfv• 7 years ago

Linux supports eBPF since 3.18.

helmettoday2• 7 years ago

right

Mathnerd314• 7 years ago

IIRC the "features" that Linux has over the BSDs are usually just referring to driver support and ported software. Since drivers are part of the kernel, BSD licensed drivers can be easily ported from a BSD to Linux, but Linux drivers are GPL and have to be rewritten for use in a *BSD.

Also FreeBSD usage is smaller so there is not as much effort to get software working on it, like Steam or obscure/specialized software. In practice Linux packaging efforts are fragmented among distros so the gap is not huge, but it's still present.

als0• 7 years ago

Not sure why Linux gets ‘Partial’ in Container Security whereas FreeBSD gets a ‘Yes’. Both provide mechanisms that can equally be misconfigured.

ti_ranger4• 7 years ago

What is the situation for "reproduce a currently-installed system from a list of packages (including version/release if relevant) and a tarball/git checkout/package-deployment/puppet run etc. of configs"?

As far as I know, on most BSDs, unless you happen to have a complete source tree snapshot for said system (becoming more difficult to maintain if you build multiple systems without special tricks to retain binary artifacts), you're out of luck, you'll get today's STABLE or RELEASE, not the one from 17 days ago when your current production box was built that works perfectly unlike the one today that is broken.

This is trivial to accomplish on all Linux distros.

(I would say this kind of issue should be listed in the table in the article).

dijit6• 7 years ago

Which BSD?

the "preferred" way to do this in a datacenter context with FreeBSD is to run a repo yourself with poudriere; this has the downside of causing you to build all your dependencies (via ports), but the upside of having everything built with your exact specifications in mind and in a way which ensures that they are rolled out in the way you prefer and that they are consistent amongt all instances.

For "old school sysadmins" who like controlling the versions of things in a very well defined way and running their own repositories it is absolutely ideal.

However, it affects developer velocity of course.

qplex• 7 years ago

It would be fair to mention that Linux actually includes two other schedulers in addition to CFQ: Deadline and Kyber.

Firewalls: Linux, and no mention of bpfilter?

darkwater• 7 years ago

It's clear that he's biased, trying to prove to the world that *BSD is better than Linux, Linux should not had the success it had etc. I remember the same people, 20 years ago too. They are probably the only type of geek worse then the typical Linux fanboy saying that Windows is bad.

joosters2• 7 years ago

The author is pretty inconsistent here. They make good points when mentioning all the problems with benchmarking and how it's a bad idea to fixate on a specific, narrow problem - but then, go on to complain about specific, narrow problems, using them to build a narrative about how linux/whatever is flawed.

A specific example: the part about how epoll() is broken because of its design, and how it causes 'thundering herd' issues when using multiple threads/processes. I had plenty of experience with epoll from over ten years ago, and it wasn't a problem even back then. Multiple processes listening on the same socket did not all wake up when a new connection came in. Sure, why not bring up specific issues with Apache, but there are plenty of other web servers that work very well (nginx, etc) and have existed for over a decade.

rachelbythebay• 7 years ago

gunicorn (Python, in case you’re lucky enough to not have run across it) forks after creating listener socket(s). Every child does the epoll thing. When awakened, they all call accept4 and then all but one gets -1.

This happens right now and “powers” a stupid amount of the economy. And it’s wasteful.

joosters2• 7 years ago

Our old mechanism was a parent process that created the listen sockets, and then sent the FDs to the previously-created child processes. (The children would be registering their FDs with epoll and then calling accept() themselves) That never seemed to trigger the problem.

pstrateman• 7 years ago

The issue is whether they're using level or edge triggering with epoll.

With level you won't miss events, with edge if you have a process crash you might.

It's a trade off.

helmettoday2• 7 years ago

I like Linux. But what about Unix?

upofadown• 7 years ago

Linux receives much more investment from companies and therefore more paid developers are in it, plus BSD’s feature parity with that of Linux doesn’t hold.

I would say that is actually true ... and is the root cause of the complexity crisis Linux is experiencing these days. There is a strong financial incentive to add things but no financial incentive to take things out. Every time you add a feature to a complex system you take something away from everything else.

johnisgood• 7 years ago

How come OpenBSD has no containers other than chroot?

Is there anything similar to firejail[1] for OpenBSD?

[1] Firejail is an easy to use SUID sandbox program that reduces the risk of security breaches by restricting the running environment of untrusted applications using Linux namespaces, seccomp-bpf and Linux capabilities.

Linux VS open source UNIX

Related Stories