OpenSolaris Indiana Released 359
Lally Singh writes "The Linux-friendly OpenSolaris Indiana has been released! A new, modern package manager and all the goodies of Solaris: ZFS, DTrace, SMF, and Xen on a LiveCD that was designed for Linux users. 'Why use the OpenSolaris OS you ask? It's pretty simple, you'll find it full of unique features like the new Image Packaging System (IPS), ZFS as the default filesystem, DTrace enabled packages for extreme observability and performance tuning, and many many more. We think you'll be quite happy to came by to take a look!'"
Who cares? (Score:4, Interesting)
Re:Still not sold (Score:5, Interesting)
There is still no mighty IOKit killer on the horizon tho... Apple (and libkern, the cpp runtine) wins.
Matt
Want to smash a harddrive like this guy (Score:5, Interesting)
http://youtube.com/watch?v=CN6iDzesEs0 [youtube.com]
zfs (Score:5, Interesting)
It's a crying shame the licensing issues keep it from being ported to Linux as part of the kernel
Re:Image Packaging System? (Score:2, Interesting)
And the stuff that isn't newly written in Java is like a throwback to the early 90's. Cryptic and hard to use. Sun uses a lot of GNU software but it's a big mix of bastardized custom stuff, stuff from the old Solaris, and GNU tools. It's difficult to get stuff working right because it doesn't work exactly like the old Solaris or something newer like Linux.
Linux kills OpenSolaris in every way.
Re:Still not sold (Score:1, Interesting)
Not yet. When home network storage servers and easy-to-use file backup become a commonplace item, however, ZFS and similar technologies will become quite important.
A pair of Apple predictions: (1) OS X will at some point in the future move to ZFS, and use its snapshotting capabilities to improve Time Machine; (2) Time Capsule will evolve into a line of user-friendly home NAS units, with ZFS under the covers, that allow users to add more storage easily, and ready for online offsite backup right out of the box (which of course, will only be compatible with .Mac).
installing now (Score:5, Interesting)
Don't take that as criticism. Cloning Ubuntu is probably the best design decision an OS team can make these days.
Personally, I don't care whether it's Solaris or Ubuntu or *BSD underneath it all, so long as it supports my hardware and runs my applications.
Re:ZFS simply rocks (Score:5, Interesting)
IP Issues with OpenSolaris? (Score:4, Interesting)
Re:Image Packaging System? (Score:2, Interesting)
It doesn't help much but it does help. It only took 48 hours to run the updates on a fresh install on my Blade (LOL, it's ridiculously slow, using the GUI version probably would have taken a solid week to finish running).
Re:Still not sold (Score:5, Interesting)
They have also forcibly crashed it over a million time and it has never lost data even once. Try doing that with your home PC.
And what
the true shame... (Score:5, Interesting)
is that ZFS, despite all its goodness, lacks some incredibly basic features compared to 99% of the hardware and software RAID and LVM systems out there. You can't grow (please pay attention here) a ZFS pool except by adding similarly-redundant vdevs, and there is no way to remove a vdev from a pool, unlike LVM2.
So. Got a 4-drive RAID-Z2 array, and you want to add more space by buying another drive to add in to your 5-bay hot-swap cage? You're shit outta luck. If you have a zpool with a vdev that consists of a pair of mirrored drives, you CAN add another vdev of two drives, then another, etc. You also CAN replace the drives in a vdev with larger drives. That's kind of half-okay, but still not on par with RAID cards of a DECADE ago. Even Linux's MD can grow RAID5/6 across more devices!
Someone suggested the ability to grow redundant pools by single devices, and the reaction amongst solaris ZFS developers (!!!) was "now why would you want to do that?", and then when THAT was explained, "well shucks, I wonder how they do that" (they = almost every hardware and software RAID solution on the planet.)
Absolutely astounding that a Solaris filesystem developer would not be able to at least guess as to how a RAID5 array would be re-striped to add a new drive.
Far as I know, they've been working on the grow capability for more than a year and we have yet to see it.
Re:Hey! It's Debian! (Score:5, Interesting)
Re:installing now (Score:3, Interesting)
More on it: http://en.wikipedia.org/wiki/Nexenta_OS [wikipedia.org]
Re:ZFS? Don't forget FreeBSD! (Score:3, Interesting)
Yes, FreeBSD has ZFS, but it's experimental for a reason. So no need to avocate this yet.
The only serious platform for ZFS yet is still Solaris, and Indiana is a welcome release.
I've also a lot of hopes in DragonflyBSD's HAMMER filesystem.
Sparc image? (Score:1, Interesting)
Re:Still not sold (Score:5, Interesting)
Sorry, I'm calling you on your B.S. Sun fanboy.
ZFS is *not* ready for production.
I'm a working Solaris admin. I can point to several ZFS raidz arrays that have had to be recovered from tape due to ZFS bugs losing & corrupting data.
This is clearly a case of ZFS marketing outstripping ZFS reality. They have implemented all the cool features, but have dropped the ball on robustness.
Do a sunsolve search for ZFS panics or ZFS corruption. There are a half-dozen major bugs that are still un-resolved, and won't be until Sol10u6 - if then. [u5 was just released in the last week or so]
rho
Re:Still not sold (Score:3, Interesting)
And what
Last time I lost data to a filesystem problem must have been to a FAT disk, which means it must have been 10 or 15 years ago. I did lose data to hardware failures though. Several times. Recovered most of them through backups. Not all.
Re:Still not sold (Score:3, Interesting)
Upon reboot two devices in the array came up with bad magic in the superblock and all was lost. The consensus seems to be that filesystem corruption caused enough confusion that the md driver decided to overwrite the superblocks. No hardware failure required, and there probably isn't enough info from the failure to find the bug. The bug is still there, and this will happen again to somebody.
Re:Still not sold - OpenSolaris in Peril (Score:4, Interesting)
Novell taking on SCO is one thing, Novell taking on Sun is quite another. Sun is a much bigger company than Novell and a lot more money. It's not worth the fight.
It seems like SCO stiffed Novell by not giving them their cut of the licenses, but that doesn't mean the licenses they gave were invalid. If that was the case, the issue would have come up already.
Novell gets some good publicity in their fight against SCO, but in reality, they're not much of a player in anything. SuSE isn't that popular, at some point their revenues for their legacy products will dry up, and then what's left? There revenue has been declining for years and their profits have been iffy. All they're going to get out of the SCO trial is some pats on the back since SCO doesn't have any more money.
While there's no arguing that what SCO did was messed up, I don't really see Novell in a good light either. Novell purchased the rights to Unix for $300mil. The transaction between Novell and SCO was for about $120-150Mill. So SCO paid about half of what Novell paid and only gets 5% in licensing fees and no patent or copyrights according to Novell.
This just doesn't seem right to me. Either Novell seriously screwed over SCO and they were too stupid to know it, or something else is going on. Ray Noorda, who was CEO of Novell, left to start Caldera. Noorda is undeniably the reason Novell was who they were. From what I could gather they did have a good relationship.
Bottom line, I don't understand how Novell can claim they pretty much just sold a 5% commission deal for 50% of what they paid and act like their shit doesn't stink either.
According the wikipedia [wikipedia.org]
Sorry but it just seems fishy to me. How would Novell not expect that SCO/Caldera would ultimately sue. Maybe Novell was aware of a possible lawsuit to attack RedHat while they were making moves with SuSE?
Re:Sparc image? (Score:2, Interesting)
Re:Still not sold - OpenSolaris in Peril (Score:4, Interesting)
I imagine that the folks at Sun have been pretty nervous since last August. Imagine, paying millions of dollars to put your product in exactly the position you've been (erroneously) proclaiming your competition is in. Not smart.
Re:Hey! It's Debian! (Score:3, Interesting)
My only real interest in Solaris is to use ZFS on a home NAS - having all that checksumming looks a lot more attractive now that disk sizes are getting so huge that, according to some, RAID 5 will stop being useful in 2009, due to the scenario of one disk failing and another one having an unrecoverable read error (URE) during the rebuild - see http://blogs.zdnet.com/storage/?p=162 [zdnet.com]. Without proactive scanning of the disk media for read errors before any failure, and checksumming that can hopefully correct some such errors, RAID 5 rebuilds after failed disks will increasingly fail due to UREs. See http://www.nber.org/sys-admin/linux-nas-raid.html [nber.org] as well for a much more technical view of the issues with RAID 5.
What if... (Score:4, Interesting)
It's designed by (deb)Ian Murdock, with 15 years of hindsight.
Re:Still not sold (Score:3, Interesting)
Are you using content of any sort (images, documents, mp3s...)? Do you care about the longevity or integrity of any of your data? Have you ever lost data? Slap a GUI on ZFS, call it "time machine" and you don't have to be "managing servers" to appreciate what ZFS can provide to Joe user.
Dtrace doesnt offer me anything as im not a developer
If you think dtrace is just for developers, you don't understand dtrace. Developers have always been able to scatter printfs in the code and set breakpoints in debuggers to glean what their piece of code is doing. The beauty of dtrace is that it allows anyone (including users, service technicians and even managers), to see what is happening on a system-wide basis without compiling or even having access to the any of the source code of processes running on that system. dtrace will prove itself far more useful for system administration and post-deployment troubleshooting than it will for developers.
But what about Joe user? Have you ever encountered an application hang? Have you ever encountered a sluggish system where "top" shows and "top" as the top process at only 3% CPU usage? Have you ever wondered whether any of your applications (e.g. spyware) are doing something they shouldn't be doing? Have you ever wondered why doubling CPU clock frequency hasn't made your environment any faster, and whether it is possible that other bottlenecks (I/O, paging activity...) might be responsible for a performance problem? Dtrace is the tool which allows you to see all of this.
SMF doesnt offer me anything i cant do with startup
Oh come on. I supported GNU/Linux long enough to know what a kludge rc.* startup scripts are. I know how non-deterministic it is and how many race conditions exist in a typical startup which work most of the time only by chance. While this is fine for a single user laptop, for enterprise looking for 6 figure reliability on massive clusters, rc.* doesn't cut it. There is a very good reason why GNU/Linux is moving towards what Solaris already has.
IPS doesnt seam any better than deb or rpm
Well, I suppose you could say the same for other *nix features: "Rsync doesn't seam(sic) any better than cp..." "tar doesn't seem any better than cpio..."
The biggest advantage for the enterprise user is that IPS is backwards compatible with the thousands of existing SVR4 packages and patches. RPM and Debian APT never will be. RPM and Debian APT do not know about zones, nor know how to handle package installs and upgrades across zones nor will they be able to ever take advantage of zfs snapshot/rollback features for seamless install/uninstall.