The sixth development release of the OpenSolaris-based, desktop-oriented Nexenta OS has been released: “Nexenta OS Alpha 6 is now available. Release Highlights: Nexenta Zones – opens the possibility to create custom zones and pre-install with selected software; integrated BrandZ – allows running Linux userland; SVR4 packaging – to install native Solaris packages (missing drivers, etc.); OpenSolaris build 50, with numerous kernel fixes and features.”
These guys have done a nice job with the FAQ. I hope some of these good documenting habits rub off on Sun 😎
http://www.gnusolaris.org/gswiki/FAQ#head-4dab4660718fa87c1f2369a44…
… on Nexenta can be traced back to Debian/Ubuntu.
Great job, people of Nexenta! I used Debian waaaaay back when, and recently switched from FBSD to Solaris. I can only hope some of the hard work you guys are putting into package management/ease of use rubs off on Sun.
Keep up the great work, it’s very impressive to see how far you’ve come in such a short time!
Nexenta is solving one of the two major complaints I have with Solaris. The first being easy package management. While blastwave is fine, it doesn’t have anywhere near the number of packages that Nexenta has in their database.
The second complaint I have with Solaris is the lack of SATA support for common chips such as Silicon Image and VIA chips. I do realize that these chips are not common in servers but its the only thing stopping me from running Solaris =(.
The webpage states that Nexenta has some drivers that are not part of Solaris or OpenSolaris, I’m crossing my fingers for some SATA drivers =)
The second complaint I have with Solaris is the lack of SATA support for common chips such as Silicon Image and VIA chips. I do realize that these chips are not common in servers but its the only thing stopping me from running Solaris =(.
The webpage states that Nexenta has some drivers that are not part of Solaris or OpenSolaris, I’m crossing my fingers for some SATA drivers =)
If you can set PATA emulation, just about any SATA chipset is supported. Otherwise, as far as I know, Solaris is still limited to Silicon Image 3132/3124 and Marvell 88SX.
I was looking into trying to set up PATA emulation. I do not believe that my BIOS supports that (Abit KV8-Max 3 mobo). Anyways, I appreciate the help Jake, thank you.
I tried Nexenta and it felt exactly like Ubuntu but without support for my printer, scanner, NTFS Windows drive, or 2D/3D acceleration from my old ATI card. Additionally, Wine does not work nearly as well on Nexenta as it does on Linux and FreeBSD using the same version let alone newer. Plus there isnt anything like Crossover or Cedega available.
While a certified Unix kernel with a stable driver ABI is no doubt intriguing, there really isnt much benefit when so few drivers are developed for it. I also think a stable ABI can impede improvement because of the need to sustain backwards compatability.
You think having a stable will encourage more vendors to create drivers for it? Better think again. Writing drivers is difficult on any platform. Vendors clearly wont divert resources into this unless there is enough demand. Period.
Besides open source drivers usually arnt affected by this unstable ABI because they can be accepted and maintained by hackers within the kernel tree. They would be included in every kernel release.
Despite the percieved lack of 1st class support for proprietary drivers, Linux is a kernel of greater innovation and supports more devices then all other Unix and Unix-like kernels.
Just look at whats happening in just a few point releases:
2.6.17 gained support for a large list of Broadcom wireless adapters.
2.6.18 has gained real-time extentions.
2.6.19 has gained code for the new ext4 filesystem.
Am I being overally cynical? I dont think so. I wish the Nexenta team good luck and I’ll try every release to see how things are progressing. But I think the NexentaOS folks exaggerate to much about their technology being superior to Linux when it clearly is not true.
Edited 2006-10-19 04:43
> You think having a stable will encourage more vendors
> to create drivers for it? Better think again. Writing
> drivers is difficult on any platform. Vendors clearly
> wont divert resources into this unless there is enough
> demand. Period.
NexentaOS using OpenSolaris core which is literally a core for Solaris and other OpenSolaris-based distributions. This means, any driver written for Solaris/OpenSolaris will happily work for NexentaOS.
> But I think the NexentaOS folks exaggerate to much
> about their technology being superior to Linux when
> it clearly is not true.
I don’t think so. Linux needs major surgery and a lot of efforts before it will reach OpenSolaris design wise. While OpenSolaris needs to close the gap on Desktop and missing drivers, which is way easier goal to achive.
Erast, I have tried Nexenta myself and think it is a very good effort and a very good OS. However, I still don’t see the benefit. So far there is really nothing that could e.g. convince me to switch servers to (Open)Solaris. Debian scales well, is going to have OpenVZ support in Etch, has two very good journaling filesystems etc. Suits me just fine. zfs, zones and dtrace are great – but it’s not enough to just add some great technical innovations
On the desktop Nexenta/Solaris is a good effort – not more, not less. There are just too many drivers missing. If that’s so easy to work around, remains to be discussed. I honestly doubt it.
The community around Nexenta is very, very small – and the community around OpenSolaris is, well, just sufficient. It cannot be compared at all to e.g. Debian’s or, on a broader base, Linux’.
“major surgery” is a bit over the top. Certainly, there are many things at the kernel level that are constantly being worked at at the kernel level. But then an enterprise customer normally has a Red Hat or SUSE enterprise edition from the start. And home users probably stick with Ubuntu 6.06 – not a bad choice.
Well personaly I agree with you.
This just goes to show that technologicly inferior products tend to get used most. I think we could find alot of comparisons in history starting with VCR vs Betamax etc.
Linux is without argument much inferior by design (I’m talking about kernels here ok?) compared to eg: freeBSD or solaris.
The problem is, Linux has “hype” and years of contributions to drivers. Solaris, altho based on old system, is “young” in this open regard, and freeBSD has smaller following mostly (as I hear) because of certain old legal issues.
I’m sure that both would be better, should they ride the same hypewave, than linux by now.
That might well be. The end user, however, doesn’t care about kernels at all. He wants a stable system with a lot of software, compatibility, speed etc. That’s what he gets when installing Linux.
And that, BTW, is what you get when you’re installing a server. So you really don’t need “a superior kernel” – at least not 95% of the time. The system has to be stable and it needs to able to scale well.
Call it what you want. Linux is backed by the major players – and that’s enough. It really is not important whether SUN is behind or not behind Linux as long as IBM, HP, Novell, SGI etc are backing Linux.
“Linux is without argument much inferior by design (I’m talking about kernels here ok?) compared to eg: freeBSD or solaris.”
The Linux design is “no design”. But I wouldn’t say FreeBSD or Solaris are much superior – they still have a lot of the same problems as Linux as far as kernel design goes.
OK, so what kernel doesn’t?
If you can set PATA emulation, just about any SATA chipset is supported. Otherwise, as far as I know, Solaris is still limited to Silicon Image 3132/3124 and Marvell 88SX.
I was not able to get a CompUSA/Silicon Image controller working despite tech support from Sun. My impression was that while a few controllers might work, this needs more testing and support from Sun.