Requested: Gentoo version

Status
Not open for further replies.
I'm very happy to make reasonable (for me) tradeoffs of time and low-level control.
I want DA to work on gentoo not only becouse of low-level control, i want it becouse of the compiling optimization becous of the emerge tool, becouse it's brilliant,
and about the time, it is not that much of time, only the 1st time when you install it, and again there are shortcats like some ppl said in here with livecds and so on.
gentoo is STABLE as hell, and works very fast becouse of the optimization.
 
jlasman said:
I hesitate to answer you Micro, because I have no idea whether DA has an unlimited budget or not.

But I know I'm struggling with growth right now, and I know that throwing money and people at a problem doesn't always fix it.

Usually as you hire people (especially when you're a small startup hiring your first people) you end up spending a lot more time on training them to do things your way than you save by using them. At least at the beginning.

And then there's the financial issue. If you're a true entrepreneur, then nothing is worth giving up control of your dream, of your company, just to get the money you need to grow.

I've been an entrepreneur most of my life, and have many friends who are also entrepreneurs. It's been our experience that when you either start bringing in venture capital our go public you lose control over your company.

Don't forget that a public company (which is where you eventually go if you want to have unlimited money) is required by law to care more about benefits to stockholders than about benefits to customers.

So I like slow steady growth myself, which often means you don't throw more money and more staff at problems.

Jeff

every company needs to advance, open more services grow.
i dont think that Gentoo is what's gonna make DA rich/poor.
 
Most DA users are not unix administrators.

Then that's something that *should* change. Why? Because people who don't know at least the basics, won't get very far with out trial and error. What's wrong with trial and error then? Why spoon feed the people that don't know how to install a program?

I'd rather DA say "Da requires this, this, and this. Here's the urls to download, install, and obtain FAQ about them." Imagine how much more time would be open to resources such as making the panel faster! :)

The other install option for DA would be for DA to check first to make sure that (a) all the required services were installed...

This feature in DA's install script apparently doesn't work all the time. I previously did an install on FreeBSD 5.x and while already installing every service by port (like it should be), it downloaded and attempted to copile and install them again.... A HUGE NIGHTMARE!!!!

That's a religious argument. I hereby request that nobody turn this thread into a religious argument about which OS to use; if it happens I'd probably delete the entire thread.

Hahaha. That made me laugh too hard. You really should try and get out more dude.

That said, DirectAdmin does things it's own way, and was originally designed to work in a Red Hat Linux environment. Everything else is a port.

<snip>

The DA staff put a lot of time into CustomApache as their way to keep DA up to date and still allow all of us to have our own custom changes. I don't see how eliminating customapache is a service to the DA community.

Everything *Should* be a port. (port not being compiled binary aka rpm, but compiled source in a source tree on the OS itself) Why is the DA Staff spending time on making a custom apache work when that has nothing to do with the control panel? The only thing apache does is work as a "service" that DA ulitizes. I see a lot of support to get DA to work with such and such.... what about the DA staff working on the control panel it self.... features, speed, etc.

This is another religious argument. How many OS's have you run DA on? Unless you disclose that, your argument is meaningless. For example, we've found DA to be rocksolid on RHEL/CentOS/WBEL.

I would *NEVER* say DA is/was solid on any OS. I would say how ever that the system itself is SOLID. You read too much into text thats not there. Please read carefully!

I have personally worked with DA on Freebsd 4.x, 5.x, and Linux Fedora. Gentoo is my current trusted OS. Even over FreeBSD.

I'd strongly urge that everyone use an officially supported OS for DirectAdmin unless the user is prepared to manage all her/his support needs.

What good is support if theres no good platforms for users to ulitize? It's the chicken and the egg theory.

That said, I'm an advanced unix/linux admin (I've administered unix systems since the 80s and linux systems since 1995, before the kernel version 1.0 release), and I'm very happy to make reasonable (for me) tradeoffs of time and low-level control.

Jeff, you might be an experienced unix admin, but you have an attitude that doesn't fly in a true Unix Admin environment. (Refering to your non-open mind and strong grudge for everything that could be good - aka Mr Scrooge.) You don't want to take chances and seem to always hide on line.

every company needs to advance, open more services grow.
i dont think that Gentoo is what's gonna make DA rich/poor.

I totally agree! If anything, the support of Gentoo by DA will not only bring more customers, but will provide a lot more $ for DA to advance on.
 
in fact i think that DA will be the 1st panel that supports gentoo ever, and this is a big + for DA.
 
Finally.... DA will work on a gentoo one.

Ok ok we get it. :) We'll work the bugs out of Debian, get the Enterprise
4.0 version out (should be quick for that one), then we can look at Gentoo.

Mark
 
Ok would love DA on gentoo, but just in case you don't know cpanel have a version working on gentoo, still have issues with frontpage but never the less it is getting close to release or so is said on cpanel forums.
 
Thanks! I did try looking on cpanel but yeah I knew they didnt have one out yet. Looks like it will be a race to see which one releases first.
 
Are we really in a race to see how many distributions we can support?

I'd much rather we be in a race to see how bugfree and feature rich our hosting panel of choice can be.

Jeff
 
I *really* don't see any benefit from having a gentoo portage of DA.

(I ll develop why later have no time right now but I had to say it)
 
The best course of action (in my opinion), to directadmin. Would be to stop adding more ports to other operating systems and focus on the 64bit and dual-core platforms. 64bit and dual-core is here to stay, especially now that intel has hit a brick wall mhz wise making AMD no longer an underdog since their lower speed (mhz wise) cpu's run just as well (and better) than intels equivalent.

The last report I heard was that AMD is no longer working on single-core cpu's (or minor upgrades if they do) and intel has nowhere to go but that route now aswell.
Sure it's cheaper for hosting businesses to just buy two servers but that'll change within the next year or two dropping down the price tag of around $1000 on dual-core processors to a more obtainable, mainstream level.

64bit is here to stay, lets make use of it and not be left in the dirt.
 
Last edited:
My $.02

I am new to DA and only mess with this on a personal level (this is a hobby for me) I have worked with several control panels and like DA thus far. I use DA on FreeBSD and am trying to understand what would be the great gain to use it on Gentoo. Why would I want Gentoo over FreeBSD?

I am a UNIX Admin for a large Fortune 50 company. One of our major concerns lately was the job of having to add Linux (RHEL) to our list of OS. Dont get me wrong I like Linux as a home user and hobbiest. I have a Gentoo box and Centos Box at home. But if you are a UNIX admin in a corporate enviroment Linux is a headache in that you have to patch constantly. Try patching over 400 machines and clusters every week. I do an emerge update && emerge -upD world once a week and it always returns junk. Now what server can be doing recompiles like that all the time?

Under FreeBSD in your supfile if you use something like *default release=cvs tag=RELENG_5_3. You will be amazed at how seldom you have to rebuild and install your world and kernel. It all just works.

As a UNIX admin I know how difficult it is to support multiple OS, I support AIX/Solaris/HPUX and now RHEL. I would assume this would be worse for a developer to have to maintain code for all these different types.

Why not in the true spirit of Gentoo (which I believe is for enthusiasts not Servers) write your own control panel that sits atop your custom OS if you need everything so customized. This way you will not be blocked by anyone or anything.

P.S. customizing any ebuild and placing them in /usr/local/portage is not different than editing any Makefile in /usr/ports
 
I've used gentoo in my server machines for some time now. I would definitely disagree that it's not a server OS. As I've mentioned before, it's not difficult to lock gentoo into certain stable versions for the things that are critical. This is actually what makes Gentoo such a great server OS... I can do a base install, drop my custom files into /etc/portage and then be assured that all my machines are running the same versions and that only the specific programs and libraries that I want installed are actually on the box. I'm a big fan of RH/Centos, but the first thing I do after a new CD install is to rpm -e all the cruft I don't need.

I'm not a BSD guy, so I will agree that setting a variable to ensure a "stable" system is probably much easier...

That said, doing weekly unattended world updates is definitely a bad idea if you want a stable machine. However, if you use pacakage.mask this can be helped greatly. Other tools like glsa-check make it much easier to never have to do a world update, while if you do need to do a big update (php, e.g.) the ebuild will take care of ensuring you have current (i.e. compatible) libraries, etc. and you don't get stuck in rpm dependency hell.

The main reason I would like to see a gentoo port is for this exact reason. There is so much tested software in portage that it becomes unnecessary to go to other rpm repositories, build your own rpms, convert tarballs to rpms, etc. This is by far the biggest pain to me when trying to extend the functionality of my rpm-based systems. For some of the scripts and plugins I've done for DA, I have to supply my own versions of the RPMs to get the complete functionality. Under gentoo, I could just, e.g., USE=mysql emerge exim.

At this point, I do agree with jmstacey, though... I'd rather see a supported RH/Centos 4 x86_64 build than another OS port. That being said, I've run 64-bit gentoo for over a year now. It would likely be straightforward to take a RH 64-bit DA binary and roll ebuilds that would be compatible with it.

BTW... creating a portdir overlay and putting your custom ebuilds there and then using package.mask to lock the system to that ebuild is very different than editing "official" ebuilds... I have no idea if ports supports this kind of functionality.
 
That said, doing weekly unattended world updates is definitely a bad idea if you want a stable machine.

If you notice I never mentioned to do an unattended world update. I said an emerge -upD world which would be emerge --pretend --update --deep world. Only an idiot would update his server without totally understanding what he is doing. This includes reading through the Makefile or ebuild.

BTW... creating a portdir overlay and putting your custom ebuilds there and then using package.mask to lock the system to that ebuild is very different than editing "official" ebuilds... I have no idea if ports supports this kind of functionality.

Yes ports supports local directories. Since portage is based on ports you will see most of the functionality. In order to better understand portage you should get to know ports.

This is an age old debate that I dont want to get into. I would just like to understand what anyone thinks they can do on Gentoo that they cannot do on some other OS
 
I'm sure I could do some of the same things in BSD that I can do under Gentoo, but I don't admin BSD boxes. Like you said, it's difficult to support multiple OSes and I have a hard enough time keeping up with the Linux distros I do support (not to mention Solaris, AIX, HPUX and Windows). I have no desire to learn ports and I'm pretty comfortable with portage, thanks.

So, for me, it's not a question of what can or can't be done... It's a question about how easy it is to do certain things and how stable and maintainable the resulting systems are. I think Gentoo and portage is a better solution than rpm-based distributions. Since BSD is not an option for me, I'm only comparing against other Linux distributions that don't benefit from the advantage of portage (or ports). I certainly didn't mean to suggest that I care about the linux vs. bsd wars.
 
I don't see gentoo as a big plus for directadmin. Gentoo is not fit for the server enviroment. Its far to unstable, I've used it on my desktop for nearly 2 years now and its been great, but its still not a server OS. Read the gentoo forums, alot of problems, I've seen a simple "emerge --update" wreck a machine. Why would anyone take a risk like this when it comes to their business?

Portage is not an aurgument as for a server os, apps like yum work fine (for a desktop then portage is a nice feature). And if its because of optimization forget it, the little bit of speed you can get is not worth the time or effort.
 
Last edited:
And you get precious little optimization because all the important programs are already compiled from source by DA.

Jeff
 
Status
Not open for further replies.
Back
Top