Why don't businesses use Gentoo?

[NTU]

Seriously, I don't get it. Gentoo is great. People talk about compiling times but if you're a multi-million dollar corporation, you lose the cost of a Threadripper in the couch. The Gentoo systems I build are sure as hell more stable than the guys at Red Hat and Debian come up with.

[Jaglover]

Why you think they don't.
_________________
My Gentoo installation notes.
Please learn how to denote units correctly!

[NTU]

[NeddySeagoon]

NTU,

Businesses do use Gentoo.

Google built ChromeOS using portage. Thats one major publicly known one.
There have been a few write ups over the years from businesses with major Gentoo deployments.
One famous one was the New York stock exchange share dealing system.

I'm sure that there are others who don't want it to be public too.
_________________
Regards,

NeddySeagoon

Computer users fall into two groups:-
those that do backups
those that have never had a hard drive fail.

[szatox]

NTU, there is a balance between "do it right" and "do it right now".
Gentoo has a steep learning curve and maintaining it efficiently would require doing some research and developing proper tools and processes, and managers and accountants don't like investments. AFAIK the industry standard is 500 linux machines per 1 administrator, which translates to ~20 minutes per server, monthly.

Now, any "standard" distro does a pretty good job when it comes to "standard" tasks like hosting websites and applications.
Gentoo trully shines when you're going for something really fancy... And once you have a bunch of gentoos around, you actually need people able to handle this creature. So you actually need people with deeper knowledge than more standard (less flexible) distros. So, basically you need expensive people who will build tools cutting costs on expensive maintenance, and then you have to maintain both, your servers and your tools, which again requires manpower and specific skills.

So... Being great from technical point of view is one thing, being a nightmare from business's perspective is another (complexity/cost and continuity/risk).

[The Doctor]

Another thing to consider is the level of support that is available. Many enterprise systems have tech support that will push a hot update for security fixes hourly if needed. Gentoo simply can't compete since we don't spend thousands of dollars for that level of support. If your a small company your probably most concerned with "it just works" rather than "it works perfectly." I knew a guy that did that for a living, although he was the guy who would drive out and do technical things, like plug the computer back in after the vacuum ran over the cord.

One could consider that my install is a commercial one with some personal use. I use it to write some documents and track business related stuff. I bet many forum members could make similar claims.
_________________
First things first, but not necessarily in that order.

Apologies if I take a while to respond. I'm currently working on the dematerialization circuit for my blue box.

[Zucca]

[Jaglover]

They dropped Windows because of latency. The transfer done one millisecond earlier beats the next one. And everything that comes after. They wanted the least latency possible - Gentoo is the winner.
_________________
My Gentoo installation notes.
Please learn how to denote units correctly!

[1clue]

While I'm 100% positive that Gentoo has a significant place in business applications, I can think of some scenarios where Gentoo would be more difficult.

Systems which must be certified free of known vulnerabilities would be one such scenario. For example, systems which handle bank transactions or process credit cards.

Those systems are generally built on some "long-term support" model, meaning old kernels and old versions of software with only bug and vulnerability patches applied. Having dealt with those systems, it seems that common sense is sometimes lost in favor of ass covering. RHEL or Ubuntu LTS or Suse are known quantities with lots of organized testing, and that sort of thing is hard to duplicate on your custom-for-the-purpose Gentoo system.

[Muso]

[pjp]

[Goverp]

I can think of two use cases:

a) A build for a generic (say) AMD-64 architecture machine, with a pretty inclusive range of USE flags, and a supporting BINHOST setup. That would build a system that could work on a variety of common hardware and offer a wide range of software for different applications. But, hang on, that's just another Linux distro like Ubuntu or whatever. I would thing the main advantage is you could eliminate certain packages (say Adobe Flash) for security or "corporate" reasons.

b) A build tailored for more specific hardware and with stripped-down USE flags to get performance/smaller attack surface/more stable environment for a particular application area. I'd guess that's why Google use Gentoo - a good platform for something like search engines.

I'd guess (a) corresponds to most businesses, who view IT support as an overhead to be minimized, and buying a RedHat licence makes sense.

(b) is for businesses who view IT as a competitive advantage to be maximised and exploited.
_________________
Greybeard

[Verdazil]

If I assume to use Gentoo in my сompany (Food industry, ~600 employees, ~250 computers), I know that I will receive such an answer: the advantages of this distribution will not fit the effort and cost of its use. Therefore, in my case - for personal use only ...
_________________
GA-Z170X-UD3 / i7-6700K / DDR4 32GB / Radeon RX 570 / TL-WDN4800 / Samsung SSD 850 EVO 250 Gb + WD Green WDC 2 Tb / BenQ BL2711U + LG TV 42LF650V

[guitou]

Hello.

I suppose the main concern is what people in business call ROI (return on investment):
Gentoo is a big investment to begin with and may offer benefit on the long term for sure, while a mainstream distro means faster ROI, lesser "risk", and so on... In an business context, short term considerations almost always prevail.
And there is a matter of "gregariousness" too: as mostly used distros will be Debian (for robustness) or Red Hat (for support), most of the choices to be made will simply avoid considering any other solution.

Even I, despite my love for Gentoo, did install Debian on my latest (and current computer), because I use it for work and happened to work in a Debian environment.

++
Gi)

[Marcih]

My workplace (a university) has an NTP and a database server running Gentoo. I was shocked when I found out.
_________________

[erm67]

[Zucca]

[Jaglover]

Custom kernel patches are modification to any OS.
_________________
My Gentoo installation notes.
Please learn how to denote units correctly!

[Syl20]

[bghoons]

I work in a mission-critical production environment. The client expects, nay, demands, 99% (I think it's 99.7%) system availability because their revenue is derived from the system being up and available to clients 24x7x365.

The client does not want to hear, "Several hundred thousand dollars in revenue was lost because...well, we had to compile an emergency patch and there was a missing dependency...uh....sorry?" No, they do not want that under any circumstance. They do not want bleeding edge, they want a super-stable system while at the same time expending the least amount of revenue to maintain the "golden goose laying golden eggs". Their optimum situation: 100% system availability with zero money spent on computing costs in order to obtain 100% system availability. If slavery were legal they would not hesitate to use slave labor, but they would not call it that, they would call it 'enforced anti-laziness work-ethic motivational system' or some other vague term to make people feel like they have worth when the client sees them as nothing more than a necessary component to doing business.

Since a free computing platform is not possible, the client retains a minimum but highly paid staff to maintain the computer system. There is Unix for legacy servers and some others as well, and there are Red Hat servers although they are the minority but seem to be on the increase as legacy is slowly rolled over to newer platforms. All of these OS vendors, including Oracle for the databases, provide a level of paid support so if there is a problem with the OS or other software, the vendor is on the hook, via contract, to fix that problem within a certain amount of time or financial penalties apply and heads start rolling.

Because Gentoo requires an extra layer of expertise (ebuilds, portage etc) beyond normal system administrator skills, as well as requiring extra computing resources to compile and distribute software, Gentoo is not viable for such an environment because Gentoo, nor any company using Gentoo that I know, has a support contract to guarantee system availability. There may be some out there, quietly doing their thing, I don't know.

The only way I can see Gentoo being more prolific commercially is for a company to hire and train a dedicated support staff of Gentoo-centric developers, software engineers, and administrators to farm out their expertise at nearly the same cost or cheaper than what Red Hat or other vendors can provide.

I would love to see that but "cheap and quick and proven reliability" is the norm in business so that is what is used. A "Great Gentoo Leader" will need to step forth and provide sound reasons why using Gentoo is the way to go.

This is just my opinion, but I cannot see myself standing before the IT Overlord saying, "You should use Gentoo instead of Red Hat, but we'd have to hire extra people to do so and it would be cool and we might even see some performance gain...." No, I cannot see it.

In summary: Gentoo would have to be like Red Hat with paid, on-demand support at the snap of the fingers, with secure repositories where all code is screened and passed on as "clean".

This is just my opinion, of course.

[Ant P.]

[bghoons]

[Hu]

[Tony0945]

A better question "Why do business users use Windows, even demand nothing but Windows?"
Why does the US Postal Service beg and plead with Microsoft to sell them NT 4.0 licenses?
I can answer that last question: Because they have machine control systems that are tied into NT. I'm sure if I had the Source code, I could upgrade it to Win 2000 or XP.
The whole system could be switched from Windoze running on two ancient computers with mobos no longer supported using Western Digital 40G IDE hard drives to a single modern platform running Linux but they don't think employees know anything.

Desktops just moved a few years ago from 2000 to Win 7.

They do use Linux for the intranet and servers - Fedora. Not RHEL, but Fedora.

[psycho]

As someone who provided user support and some UNIX and Windows server admin for a large organisation for many years, I would be reasonably happy to use Gentoo for a personal desktop for work stuff, or to recommend it to power users, but I would never deploy it enterprise-wide for standard user desktops: to me that just seems ridiculous. There are good stable distros with long term support that you can depend upon to work the same way day after day, year after year: you can develop all your training materials and so on, knowing that the instructions are going to work precisely the same way for everyone for the next couple of years (or whatever) before they need to be updated... you just have a much easier life for a whole host of reasons, all of which would go straight down the toilet if you deployed a source-based rolling-release customisation-focused distro as a standard rollout for normal users.

If GLEP 19 or something along those lines (a frozen, maintained-with-fixes-only Portage tree) were ever implemented, Gentoo could be a viable business/corporate distro... in fact I'd say its ability to do both that *and* everything it currently does would make it almost objectively "the best distro", if it gained a large enough user base, as it could meet almost any need better than almost any other distro... but this just doesn't interest most Gentoo users: folk who want the dependability of a stable (as in not-changing), standardised (maximum use of identical setups for the maximum odds of catching the bugs on that setup), LTS distro already have several to choose from, and folk who want a meta-distro that empowers us to do all the customisation that Gentoo can do tend not to be particularly interested in using all that power just to re-create another stable OS like Red Hat or Ubuntu LTS. Gentoo's great for developers, power users, hobbyists, researchers and so on... but when it comes to business deployments it's just not the best OS for the job: I don't see how anyone who's experienced the realities of large-scale IT administration could have any trouble understanding why those deployments are done using Red Hat and Ubuntu and so on. Change and variation, the very things Gentoo does best, are your worst nightmare when you're trying to keep thousands of (often fairly incompetent) users happy without losing your mind.

Gentoo's a niche, power-user distro that could expand and become every granny's and suit's distro if it wanted to...but it doesn't want to, and I doubt that will ever change. We'd need a significant distro-wide developer commitment to maintain a frozen portage tree in addition to everything else, and I think most of the people motivated to maintain a stable OS like that are likely to be doing that already, rather than here using Gentoo. Gentoo is not an "enterprise" distro: I've often seen people in these forums claiming that it is, but it just isn't. The great enterprise distros are the ones that focus on the enterprise, keeping everything predictable and standard: Gentoo's focus is very much in the other direction, on supporting every individual user's making it into whatever we want it to be.