OSNews

He really did say anything Ubuntu based was out of the question.

I myself have set up I nice HTPC with Arch Linux http://www.archlinux.org/ though it might be more work to set it up then Ubuntu or other more graphical based distrobuttions.

For read/write support with ntfs partitions check out ntfs-3g http://www.ntfs-3g.org/ (If your going for Arch Linux; it's available in there repositories.)

Boxee is available in Arch'es AUR http://aur.archlinux.org/packages.php?ID=24206

I'm planning on building an Arch Linux based media centre too. I think that Arch is perfect for these types of job (though I am a committed Arch fan, full disclosure etc.) since you can easily custom-build a lightweight, up-to-date system from scratch but I probably would hesitate to recommend it to someone who isn't very familiar with Linux in general. You will be expected to use the command line, edit text files and all that stuff.

I'm actually leaning towards Mandriva too. I'm going to figure out if it makes it easy to install Boxee and Chrome.

Yes, install Mandriva, then VirtualBox and install Ubuntu 9.10 in VirtualBox and then install Boxee. Then run Boxee in Ubuntu in VirtualBox in Mandriva.
The Ubuntu problem should not exist in VirtualBox and you can still have a familiar distro with everything you like.

Seems like a little over kill of Boxee runs natively under Mandriva though. I'm also waiting for two developments in Virtualbox:

- True "bridged" network devices without scripting tunnel devices by hand. Bridged in addition to NAT networking should have been in the first draft on that one.

- Start at boot support for VMs. I run a groupware server VM guest under my workstation. I need that VM to autoboot when the host system reboots. This has to be possible without scripting and juggling between account priviledge.

Those are not impossible issues; they just keep me from being able to replace VMware with Virtualbox for my needs.

VMware Server on Debian Lenny has it's own issues related to the SSL enabled web interface. I'm able to use it an access my older VMs through the http:// interface but no https:// and the related service hangs shutdown and restart (system or VMware services).

Now, VMware Server under Mandriva never gave me issue. Should such a complicated setup be needed, VMware should play nice with your Mandriva and autostart a bridged VM for Ubuntu or one's preferred choice.

Hi, I haven't touched that in ages with VirtualBox. It just works in recent releases. I can't comment about your other problem, but bridged mode is just a down arrow away!

Hi, I haven't touched that in ages with VirtualBox. It just works in recent releases. I can't comment about your other problem, but bridged mode is just a down arrow away! "

It's called Host Networking, which might trip people up. But it's actually bridged, using the virtualbox network filter driver in the host, as opposed to a separate bridge interface.

- True "bridged" network devices without scripting tunnel devices by hand. Bridged in addition to NAT networking should have been in the first draft on that one.

- Start at boot support for VMs. I run a groupware server VM guest under my workstation. I need that VM to autoboot when the host system reboots. This has to be possible without scripting and juggling between account priviledge.

Virtual Box already does both. I have 4 VMs starting on boot on a Ubuntu Server running Virtual Box. All of them connected with bridge networking (without scripts).

Search "easyurpmi" and you should end up at the right place to add network and PLF repositories.

I second Sabayon. It implements a true BSD style ports tree where you have both pkgsrc and pkgadd except pkgsrc is Gentoo's emerge and pkgadd is equo. It's my first choice for any job.

Have you tried the alternate text installer? If not, really you should. It's not the same kernel as the liveCD one. Moreover, that the installer is crashing does not mean the system will crash afterward. Again, it's not the same kernel.

I create my USB boot things with unetbootin, but that one doesn't include the text installer option. On top of that, people are reporting the same problems with running systems - it's not just the live CD kernel that fcuks up. It's all a bit of an epic fail.

I'm kind of bummed. 64bit didn't work either.

Err, the Atom is a low end-chip. So no AMD64.
That is market segmentation for you.

Edit:Sorry, I was wrong. Only the Z and N don't have 64bit.
But with less than 4GB 64bit is toally pointless. You waste a lot of CPU cache for nothing. The system will be way slower.

Edited 2009-10-30 16:59 UTC

Strictly speaking, don't the kernel tradeoffs begin above around 800MB? Bounce buffers and such? And don't forget all the extra registers that come with X86_64.

Personally, I think that 64 bit kernel with 32 bit user space, with individual 64 bit apps deployed strategically, is really optimum for many situations. The kernel has all the memory management advantages of 64 bit. But without the memory hit of 64 bit user space pointers, except where 64 bit can really help. As it *which* 64 bit apps to strategically deploy, if any... I'm not certain. And, of course, it depends upon the exact situation.

Yes, you are right. 64bit kernel and 32 bit apps would be optimal in most cases.
AFAIK modern CPUs get their speed from running stuff in cpu cache and 64bit pointers etc take way more cache memory.

I don't know that much about CPU design, but I think the current design of X86 and AMD64 has a lot of limitations because of backwards compatibility.
Maybe when FOSS gets more popular AMD or Intel will introduce an "X86 unleashed" mode where you have can 16 64bit registers and 32 bit pointers and GCC and go wild with its optimizations. Might be just a brainfart, though.

My impression is that the main problem with x86 was its register-starved nature, and lack of 64 bit instructions. Both fixed by X86_64. The rest of the complaints about x86 seem pretty ivory tower. On a computing power per dollar basis, X86_64 looks to be sweeping pretty much all the competition either into the dirt bin, or into little niches in the corners.

Would some other instruction set be a little better in some way? Does it matter?

You don't need any changes to the processor to do this. As I understand it, a few different JVMs can operate in this mode (32-bit pointers stored in memory, plus some base offset). FWIW, Itanium actually has special instructions to make 32-bit pointer size systems fast, so Intel has clearly thought about the strategy.

Definitively give 64-bit a shot. I have an Intel board with the same chipset for a media backend, and it runs ubuntu-server-amd64 without a hitch.

You could also give the netinstall (mini.iso) a shot.

Finally, try unplugging the sound card during install. It should not make a difference, but there is a chance that that card's module is causing your problems.

Another thing to consider: If you are using your HTPC as a mythtv backend, you may want to stick to 9.04. Karmic has updated the mythtv packages to 0.22, and they are not backwards compatible.


Edit: Posted before reading the RE[2]: 64bit. Still, consider unplugging the sound card.

Also, do you have a chance of checking the error logs? It would help narrow down the issue.

Finally, I know you said it ran ok on Win7, but have you run memtest?

Edited 2009-10-30 14:58 UTC

And the main reason why I am mostly on Windows nowadays on the desktop and leave Linux on the server room.

I would gladly be on MacOS X, if they weren't so expensive in Europe.

I am a Linux user since the kernel 1.0.9 days, and nowadays have more interesting things to do with my life that spending endless hours with such configuration issues, or broken releases. My desktop should just work.

If you want a desktop to just work, you should only use stable software, man.
People use testing software, which are advertized as testing software: for tests, which means: "please help us test this software, which probably is buggy" and then they complain that it does not work!
Man, if Windows development was open and they had a portal where you could download all their testing stuff, you would probably be crying!
People complain and they don't even use the bug tracking system.
Well, maybe linux is not for you. It's open and documented but if you can't read, it won't help you.

Since when is Ubuntu 9.10 testing software? It's a full-on release, last time I checked.

It's based on Debian testing.
Canonical releases LTS versions, which are more stable. Other versions are pretty bleeding edge. When Dell chooses Ubuntu for their netbook, they pick a LTS version, because other versions are not as stable and supported by Canonical.
If you want stable software, use Debian stable, or Ubuntu 8.04 LTS.
If you want to try new things and be on the bleeding edge, use Ubuntu 9.10, but don't complain when there are bugs, post them on launchpad instead.

It's the same when you use Fedora instead of Red Hat.

Edited 2009-10-30 15:40 UTC

A release is a release. There is no testing warning or whatever on Ubuntu's site. Why are you defending such an epic fail on Ubuntu's end? I mean come on - Intel video chips causing freezes all over the place?

That's unacceptable.

I'm not defending Ubuntu in any way. Ubuntu is based on Debian testing.
Look, if you use Fedora, you know it's a bleeding edge distro and it's a testing ground for new things.
It's your choice to use it, nobody put a gun over your head. There are bugs, yes and we all know it.
Now if you don't like unstable software, use Red Hat and pay for support, or use CentOS, or Debian stable.
There are literally more than 500 distros to choose from. YOU make the choice, according to what YOU want. Now if you want stability, DON'T USE BLEEDING EDGE DISTROS!!! Slackware, Red Hat, Debian (stable), etc... are made just for you! use that and stop whining.
If you still use bleeding edge distro, don't go around saying stupid things like 'linux is buggy', 'Windows is stable', etc... you just chose to use testing software.

Edited 2009-10-30 15:55 UTC

I did not say such a thing. All I said was that if the most popular Linux distribution has major problems for the second release in a row with Intel video chips, then that's an EPIC FAIL. Bleeding edge blah blah or no, that's just totally unacceptable.

No, it's based on Debian Sid (debian unstable). They rebase their packages on Debian Sid every 6 months, and then spend a couple weeks re-integrating everything, then spend a couple week testing, and then release it. It's not that great of a release process, IMO.

Just read what linux "fanboys" just use to say: maybe linux is not for you:

http://linux.oneandoneis2.org/LNW.htm

I had Debian 6 Testing on my machine for a few weeks. Some of the additional software was great (OpenVAS) and I actually liked KDE4 now that it's had a few point releases since. Ultimately, I ended up dropping the notebook back to Debian 5 Stable though. KDE4 and the additional apps did not justify some of the breakage I was getting with updates. That's not a complaint about Debian, I knowingly ran a testing status distribution. I've since obtained OpenVAS through backports and KDE3 runs lighter on the system so I can wait for Debian 6 to go stable or at least much closer to stable release.

I think the point the previous poster is trying to make is this same thing; when you run a testing distro you have to accept that your using the beta of the next release. When you run an RC or Beta Windows release, you have to accept that it's in development and may break or lack features.

This should not excuse Canonical for so many having issues with such a generic chipset and GPU. It does indicate a potential cause and justify reverting back to an LTS, trying the parent distro Debian or another distro entirely.

(I have to admit I'm a fork snob also; why use a forked distro if the parent distro provides all the same functions and/or more.)

Bottom line; popularity has never indicated the quality of software. Canonical is popular but there are other distros more applicable to given tasks.

Thom, I agree. That is an epic fail. I'm very surprised that Ubuntu didn't catch that bug during the alpha-beta-rc cycle. As you pointed out, it's not like Intel's integrated chipset is some kind of obscure graphics chipset. How did they miss that bug?

--Johnny

How they missed this bug?

Simple: Time based releases!

If you have a release date, you are always going to have a huge number of bugs, that is simply unavoidable.

The last Debian STABLE (lenny) was released, when the number of release critical bugs was just less than 100. Now imagine how many more bugs a distro must necessarily ship with, if debian sid is taken as a base and 6 Months of bugfixing goes into the system.

You might have hardware which works like a breeze, then it's a very fine system. If you don't have this hardware, you are out of luck.

In this case it seems like buggy Intel drivers are a major cause of pain.

Am I the only one who has never had any major problems with Ubuntu on the desktop? Windows gives me headaches all the time. Various apps will crash the system, I get constantly high CPU usage that overheats my laptop, the UI seems very inconsistent (once you have apps from all over the place), and I find it to be very unintuitive compared to Ubuntu. A Mac would be nice, but they're ridiculously expensive.

And would burn your balls off on its way to your wallet, too.

Did you try the vesa driver then?

I'm waiting for the price of nVidia ION systems to come down a bit, and then building my HTPC. So far, the Acer Revo looks just about perfect.

* Dual-core Atom 330 CPU @ 1.6 GHz
* 2 GB RAM
* 160 GB HD
* nVidia 9400M graphics onboard
* HD sound onboard
* VGA and HDMI output
* several USB 2.0 ports
* gigabit NIC onboard
* wireless N chipset on board

Comes with a VESA mount, so it can be mounted on the back of monitor/flatscreen. Now to find one for less than $400 CDN.

All the reviews online show this setup (dual-core atom + nvidia 9400) capable of doing 1080p without hiccups, and only around 30% CPU usage.

Sue me.

From my following the recent discussions; isn't the EULA unlawful in Netherlands indicating that his building a hackintosh was fully within local laws? Since a post sale restriction does not apply, copyright law allows for the legally purchased content to be installed provided it's not installed in two or more places?

Agreed. But then again, OSNews isn't really famous for a professional image. The site is rife with dilettantism and arguing over the smallest trifles. Join us next week when someone needs advice on how to trick OS/2 into installing on a PowerMac G3 they found on the piss-stained floor of the boiler room at their small town high-school using floppies. Laugh as the entire OSNews gang practically murders each other when six different people come up with six different ways of accomplishing the same pointless/impossible task!

Since it obviously upsets you, I'm going to go install OS X on a non-Apple PC right now - just for spite.

Hahaha!
- BallmerKnowsBest

What law did he break? An EULA is not criminally enforceable, else half the computer-using people on earth would be rotting in jail. He didn't violate the DMCA (not that it matters in his country anyway) and he certainly didn't steal anything. If you'd read his article on the Hackintosh process he used, you'd know that. He used a Leopard retail disc, a boot-132 bootloader (open source under Apple's own license) and he even labeled the computer with an apple--tongue in cheek maybe, but to the letter of the EULA.

So, as a law enforcement employee for the past ten years, I ask you: What law did he break? My coworkers certainly wouldn't arrest him for installing legally purchased software on a legally purchased computer with no circumvention of copyrighted content. In fact, they'd likely arrest YOU for false report of a crime, which is illegal.

Really? I've always wanted to try Debian. But I could never get it to install on anything.

Installs on whatever I throw it at here. But then I don't use the graphical installer, maybe that makes a difference?