Comments on: Why Gentoo Shouldn’t be on Your Server The Internet Startup Blog Sat, 24 Oct 2009 19:07:30 +0000 hourly 1 By: riaan nolan Fri, 05 Oct 2007 09:27:31 +0000 On the contrary, I only use Gentoo for servers. It’s the most secure, stable server OS that I have ever worked with. Yes updates can brake things, but that’s why you have a development server to see what changes first .. at least you should have such a strategy in my opinion.

Gentoo Rules!
star@flyer star $ uptime
11:28:14 up 739 days, 22:30, 1 user, load average: 0.00, 0.00, 0.00

By: required Sun, 16 Sep 2007 02:19:52 +0000 i can understand your points, but i don’t think they are necessarily deal breakers to use gentoo. i feel you have a bit more control of what you are doing with gentoo, and end up with a very minimal installation of what you need.

if you do not want an upgrade, you can mask that version. granted the install may take a while, depending on your hardware. i think since 2004.0, gentoo has come a long way, and is definitely a contender.

you have a lot more control with gentoo than you are giving credit to. i think it might take sometime to know, but in the end, its very slick.

maybe it comes down to a choice.

i don’t think gentoo is an update everything. you can be mindful of what by adding ur pv flags to see what versions would change if they do.

i don’t think gentoo is necessarily bleeding edge. you could choose that route, and possibly have a broken machine. the community is rather large, and you have a dedicated group of testers, who will put up to the head arch dev/tester to add to the particular stable flag.

i’ve been a gentoo user for a very long time, and i have my gripes, but i don’t think i would choose anything else over it.

its not as broken and cumbersome as ur original article has stated it to be so.

By: kRAkEn/gORe Thu, 06 Sep 2007 09:10:20 +0000 i’ve run redhat, suse, and gentoo servers and i can assure you that updating a profile is much less error prone than update some packages in a redhat server. i’ve had more problems with redhat system failure after a global system update than gentoo.
yeah gentoo is time consuming, and it requires you to be careful. not for all the push-button servers lovers :)

By: anonymous Tue, 12 Jun 2007 01:37:46 +0000 [Deleted due to language.]

By: Gentoo | rakhesh, deblogged Thu, 17 May 2007 10:58:13 +0000 […] Not related to 2007.0 but here’s an interesting Gentoo article I read a while ago. § […]

By: Alexander Ljungberg Sat, 12 May 2007 18:55:14 +0000 I think you misunderstand what I was trying to do here. I was trying Gentoo out. For me it makes more sense to use old hardware for that, rather than taking out hardware I actually need for anything.

A lot of people have said that redundancy is the answer. This is true to some extent, but I would much rather have a system less prone to breakage with or without redundancy. Even if you have a staging server for setting things up, you’ll be much better off if things work more often than not. What’s worse, even with a staging server, you’re ‘playing with fire’ as you put it. You plain may not notice breakage the latest updates introduce before you move them from the staging server to the production server. As any software developer will tell you, you can never debug anything to a hundred percent.

If you think Gentoo requires two servers for a production environment, you would still be better off using two RedHat Entreprise servers or two FreeBSD servers. If your redundancy halves your risk, or even brings it down to a tenth, you’ll still be better off with less base risk.

This article has had several people agree and several people disagree. Therefore, even if the opinions expressed in the article were the complete opposite, there’d be people asking me to rewrite it. You’ll understand if I’ll just let it stand as is and let people draw their own conclusions based on the data presented.

By: Crustybum Sat, 12 May 2007 17:13:09 +0000 Wow, this article shows up on google when searching for various Gentoo-related things.

Unfortunately, the article smacks of ignorance and lack of forethought.

First, nobody in their right mind should be installing a source-based, compile-everything OS on old hardware. If you really need to run old hardware that is lacking in the power to be any kind of compile platform, then don’t run Gentoo, Sourcemage, or anything else that requires considerable compile time. You’re just setting yourself up for frustration, and that’s not the distribution’s fault.

Second, Gentoo has a nasty perception of being a ricer distro, but that perception doesn’t continue to need to be propogated. The Gentoo documentation doesn’t encourage people to setup psychotic compile optimizations, so anybody coming in new and just following the docs shouldn’t be attempting to do so.

Last, in any production environment, if you’re really serious about providing 100% service uptime to people, not only do you setup redundancy so that a problematic system update doesn’t actually disable service, but you also test each and every software update you do before rolling it out on a production machine – binary distro or not. If you don’t, you’re playing with fire whether you trust the OS or not. FreeBSD, RedHat, Windows, whatever, you don’t just blindly hope your updates don’t take out services.

You may want to seriously reconsider re-doing this article. There’s a lot of mistakes in your methods that lead you to come to your conclusions.

By: John Ward Sat, 12 May 2007 08:11:56 +0000 hmm.. i have 13 production servers running everything from jbos to mysql and backup admin stuff. All running on Gentoo.
At worst, i’d say it takes me 2 hours (in between meetings) to get a raid server going.
It takes a full 20 minutes to install a fresh machine off tape ;) i really do think a bit of RTFM is required ;)

By: JD Fri, 11 May 2007 16:24:25 +0000 Sadly, i sort of have to agree with the autohr on the base of the article, but i have a different point to bring up…There’s one big reason Gentoo wouldn’t be good for a server and that’s the fact that essential packages constantly updating (there’s never a “stable”) and if anything happens to bork the system, your server is down and most servers can’t afford that much downtime of having to find and fix the problem.

By: ChaosRequiem Tue, 27 Mar 2007 06:43:20 +0000 Of course another thing to consider (at least from my unconventional setup) is the bandwidth. 56k and constant updates are not pretty. I’m currently using Slackware 11 on everything. (Ironic, same OS for both network backbone/security, and graphics work.)

Hopefully if our connection gets upgraded next week due to shifting service areas… it still begs the question. Bandwidth spent updating servers is bandwidth not spent doing what they are supposed to be doing, no matter what OS/Distribution.
Touchy hardware aside. My current internal server at our little start up qualified as a toy in 2002… Good enough for dialup and internal DNS.