It shouldn't apply. ZFS isn't a network file system. If you were using a SAN you could in theory reformat portions as ZFS, but NAS wouldn't be affected unless the vendor of the NAS decided to update their unit to support ZFS internall.
Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
ZFS To Become Default File System In Leopard
- Thread starter MacRumors
- Start date
- Sort by reaction score
It shouldn't apply. ZFS isn't a network file system. If you were using a SAN you could in theory reformat portions as ZFS, but NAS wouldn't be affected unless the vendor of the NAS decided to update their unit to support ZFS internall.
Does it make any sense for the filesystem to automatically comandeer external drives without asking the user first?
Really, why do people keep asking this? The only time I can see this possibly being a risk is if you were using a eSATA connection to one of those brackets that hooks right up to the motherboard's controller.
I don't know that Finder or Spotlight would really be terribly effected but Time Machine would certainly be nicer with ZFS' integrated snapshot capability.
Does it make any sense for the filesystem to automatically comandeer external drives without asking the user first?
Really, why do people keep asking this? The only time I can see this possibly being a risk is if you were using a eSATA connection to one of those brackets that hooks right up to the motherboard's controller.
ZFS pooling of devices is not automatic, you have to tell ZFS via the command-line or some yet undisclosed GUI to expand a pool, it doesn't just happen automagically...
kiang is right. ZFS has almost limitless storage ability.
Quote from Wikipedia:
If 1,000 files were created every second, it would take about 9,000 years to reach the limit of the number of files.
Project leader Bonwick said, "Populating 128-bit file systems would exceed the quantum limits of earth-based storage. You couldn't fill a 128-bit storage pool without boiling the oceans."[1] Later he clarified:
Although we'd all like Moore's Law to continue forever, quantum mechanics imposes some fundamental limits on the computation rate and information capacity of any physical device. In particular, it has been shown that 1 kilogram of matter confined to 1 liter of space can perform at most 1051 operations per second on at most 1031 bits of information [see Seth Lloyd, "Ultimate physical limits to computation." Nature 406, 1047-1054 (2000)]. A fully populated 128-bit storage pool would contain 2128 blocks = 2137 bytes = 2140 bits; therefore the minimum mass required to hold the bits would be (2140 bits) / (1031 bits/kg) = 136 billion kg.
To operate at the 1031 bits/kg limit, however, the entire mass of the computer must be in the form of pure energy. By E=mc², the rest energy of 136 billion kg is 1.2x1028 J. The mass of the oceans is about 1.4x1021 kg. It takes about 4,000 J to raise the temperature of 1 kg of water by 1 degree Celsius, and thus about 400,000 J to heat 1 kg of water from freezing to boiling. The latent heat of vaporization adds another 2 million J/kg. Thus the energy required to boil the oceans is about 2.4x106 J/kg * 1.4x1021 kg = 3.4x1027 J. Thus, fully populating a 128-bit storage pool would, literally, require more energy than boiling the oceans.[5]
http://en.wikipedia.org/wiki/ZFS
Is there smoke coming out of Steve's ears today with this slip announcement? Or was it even a slip? Is this another RDF pre-wave for WWDC?
Right, this is just the teaser for the Big News between the two companies (no, they're not merging).
Apple has incorporated so many open source projects into OSX I'd be very surprised if FUSE didn't make it. And maybe they're fix the rename over sshfs problem...
HFS+ is old, but not mature. It melts itself all the time. Same goes for AppleRAID. ZFS has, IIRC, three known data corruption bugs.
ZFS focuses on storage, but not so much on semantics. There's room for Reiser4-style semantics in ZFS.
Go look at the Time Machine videos. You'll also have a greatly reduced chance of disk problems.
It was the obvious problems (two disks in a MacBook) that were so surprising last time around. ZFS fixes that.
No, TimeMachine needs this.
My goodness, there are so many solaris kernel features showing up in Leopard.
Yes. Whether the GUI exposes this is an open question.
Good question. If the Mini is discontinued it's because the new iMac makes it redundant.
USB can come within 20% of the speed of firewire these days. Not too shabby.
I m being admittely lazy today... and the fact that the sun presentation is VERY boring.. could someone tell me the timecode where he talks about zfs and mac os x?
It's already been posted in this thread. Be as lazy as you want, but don't expect others' to carry your water.
Copy-on-write for performance and data integrity:
Basically means that modified data doesn't overwrite original data realtime. This means more sequential writes (faster) and protection from system crashes that kill your OS install with data snapshots/checkpointing.
Data integrity through full-time checksumming:
Hard drives can corrupt data as blocks go bad, but ZFS checksums all data realtime. This is a big benefit in RAID configurations, but not sure about how much benefit for single-disk.
No more partitions:
I'm guessing this only affects ZFS partitions. 'Volumes' that grow as you need them.
Since Adobe's CS3 was just released recently and it still doesn't work on UFS file systems, or even case-sensitive HFS+ file system I doubt it would work on an entirely new filesystem. This seems to indicate that using ZFS as the default filesystem would be premature to say the least. Developers are very slow to implement new technology and for Apple to force everyone to re-design their apps would be very unwise. I personally think that ZFS should be the default filesystem on 10.6, but only an optional ability on 10.5.x
The name was officially changed to 'ZFS' a while ago.
I think you'll have to reformat. Yet another reason to hold off on buying that new Mac until after Leopard ships installed on them.
This is one of those features Windows really needs. Only documents in a user's "Documents and Settings" folder are kept private. And that folder is part of the OS filestructure. So when your boot disk fills up, you can no longer store docuements in your "area" everything you put on other internal hard drives is just there for anybody to go through.
It would be an option for new installs since using it would mean having to reformat your existing HFS+ drive (and needing someplace to store all your data in the mean time). But then, this could be one of those golden opportunities for an initial Time Machine backup.
I wouldn't be surprised if they made this an Intel only feature, too.
But then, you would need the ability to have both devices hooked up at the same time, might be a problem with the limited expandability of some Macs.
Would it be possible to place the new drive in an external case, shuffle the documents over USB or Firewire, and then take the drive out of the case and stick it inside the computer in place of the drive we're replacing and have ZFS still see everything matching up?
Yes, that was my point.
Nobody would be assuming that if they'd just think for a second before posting.http://daringfireball.net/linked/
Marc Hamilton, Suns VP for Solaris Marketing:
"Jonathan noted that Apple will announce this week that the ZFS file system from OpenSolaris will become Apple's new default file system. So how does that help Sun? It is pretty simple, now every Apple developer will know ZFS and how to use it on products like our SunFire x4500 storage server and other Sun products."
http://blogs.sun.com/marchamilton/entry/sun_s_new_modular_blade
Marc Hamilton, Suns VP for Solaris Marketing:
"Jonathan noted that Apple will announce this week that the ZFS file system from OpenSolaris will become Apple's new default file system. So how does that help Sun? It is pretty simple, now every Apple developer will know ZFS and how to use it on products like our SunFire x4500 storage server and other Sun products."
http://blogs.sun.com/marchamilton/entry/sun_s_new_modular_blade
Unless the new iMac has no built-in monitor and starts at $599 I don't see any way this statement makes any sense. Most people buy Mac Minis to be used as computers, not expensive NAS devices or a Pseudo-AppleTV.
Video?
Does anyone know of a [working] video/webcast of this? MacYang - where did you see the video???
Does anyone know of a [working] video/webcast of this? MacYang - where did you see the video???
You are right. There is no way to do this with ANY file system. It's logically imposable. But what about the case where the machine has just one drive. I think 90% of the Macs sold will have only one drive. OK make that 99%. Not covering the odd case where a user has two drives of un-equal size seems unimportant in light of all the single drive systems
My guess is that ZFS will NOT be the default file system on single user desktop systems but will be used on servers. I'm pretty sure that right now, at least, you can't boot from ZFS.
Well ZFS would be of some use even on a single drive. It has that "Time Machine" thing built into it. It can't save you from a drive crash but it can a "ops I didn't mean to delete that paragraph. Let's restore the version from last week." Also ZFS can allow a full clean "point in time" backup of a drive that is in use.
ZFS looses most of what is good about it if you only have a few drives
What is going to be real interesting is to see how Apple handles the user interface for the administrator. Sun has done a truly horrible job at this. I hope Apple has come up with a way to make the admin job simpler. THe concepts that are inside ZFS ("virtual storage pools" "transactions"...) are way over the heads of most Mac users and require Sun Admins to hit the books pretty hard if they are going to be using ZFS.
Again I doubt Apple will make ZFS the default on syystems that have only a handful of drives. ZFS is for the systems with racks of drives to manage. Our system down stairs has hundreds of fiber channel drives and supports a few hundred users. This is what ZFS was created for.
A few things: First, it is good to see changes in several filesystems that have been coming over the last few years. Even the dreaded MS has had some really good (however well implented) features that could be of real use to people. I have been in the Unix world for quite a while, and I can tell you that some of the things we have had to live with for a long time are why companies like Veritas made a lot of money. Managing disks is a thankless job, and do it effectively was not trivial. But things like journaling that increase IO by only a few %, extensive volume managemet features, and checksumming that does not make the system crawl are all really, really cool; and ZFS is one option in this new, brave world. Hell, has anyone had to fdisk a 300gb drive recently? We had one plug in to an older sparc solaris box, that did not have journaling enabled . . . well, let us just say that it tool a *long* time. And journaling is not a new concept; just one that a lot of players took their time accepting.
I have a lot of outstanding questions, of course, but after speaking with one of my colleagues, I thought I would post one and make a comment on another. First, does anyone know of a BSD implementation of ZFS? We can get a good idea of what the feature set will be from looking at Solaris, but I guarantee you that it will not all be the same. It could even be that Apple puts some thought (and it is not trivial, so it will take some thought) in to actually booting from ZFS partition, so they would actually be "ahead" of Sun in that regard. But my feeling is that since Mac OS X is nearest to some of the BSDs out there, I would think those implemented features might tell us more. I see some abortive attempts on google, and apparently ZFS is in the FreeBSD 7.0 Head branch as experimental. I am not exactly sure what that means, but . . . and apparently, not all archs are ready so . . . Anyway, maybe someone knows more.
The second . . . statement is that I am so glad that we have what appears to be a way to get around the "lowest common denominator that everyone seems to be able to use is fat" problem. I play with many architectures, and besides fat and iso9660, it is really hard to port disk around. I guess the closet might be ext2 (they have support in windows, though it is a bit clunky; but of course you loose even some of the basic features in ext2) and xfs (but I have not seen a workable implementation on windows, and its code base is not the cleanest). I guess maybe I should mention ntfs, but its linux support is . . . so, so, and Solaris . . . No, nothing really works easily for me that can handle multiple disks easily, had many of the advanced features, performs welll . . . Maybe, just maybe with two senior platform companies (Sun and Apple) pushing it . . . FUSE implementation of ZFS on Linux seems to be in some stages of use, there are some people thinking of ZFS on windows (gotta love open source) . . .
I am not naive enough to think it will be a panacea, but I can hope that I could move my 500gb usb 2.0 connected drive from a Mac to a Linux PC to a Window PC, and be able to read my "downloaded" HD DVD copy of Pirates of the Carribean . . . without having to think of nfs or smb.
I have a lot of outstanding questions, of course, but after speaking with one of my colleagues, I thought I would post one and make a comment on another. First, does anyone know of a BSD implementation of ZFS? We can get a good idea of what the feature set will be from looking at Solaris, but I guarantee you that it will not all be the same. It could even be that Apple puts some thought (and it is not trivial, so it will take some thought) in to actually booting from ZFS partition, so they would actually be "ahead" of Sun in that regard. But my feeling is that since Mac OS X is nearest to some of the BSDs out there, I would think those implemented features might tell us more. I see some abortive attempts on google, and apparently ZFS is in the FreeBSD 7.0 Head branch as experimental. I am not exactly sure what that means, but . . . and apparently, not all archs are ready so . . . Anyway, maybe someone knows more.
The second . . . statement is that I am so glad that we have what appears to be a way to get around the "lowest common denominator that everyone seems to be able to use is fat" problem. I play with many architectures, and besides fat and iso9660, it is really hard to port disk around. I guess the closet might be ext2 (they have support in windows, though it is a bit clunky; but of course you loose even some of the basic features in ext2) and xfs (but I have not seen a workable implementation on windows, and its code base is not the cleanest). I guess maybe I should mention ntfs, but its linux support is . . . so, so, and Solaris . . . No, nothing really works easily for me that can handle multiple disks easily, had many of the advanced features, performs welll . . . Maybe, just maybe with two senior platform companies (Sun and Apple) pushing it . . . FUSE implementation of ZFS on Linux seems to be in some stages of use, there are some people thinking of ZFS on windows (gotta love open source) . . .
I am not naive enough to think it will be a panacea, but I can hope that I could move my 500gb usb 2.0 connected drive from a Mac to a Linux PC to a Window PC, and be able to read my "downloaded" HD DVD copy of Pirates of the Carribean . . . without having to think of nfs or smb.
I'm interested, but as somebody with a number of Mac's I'm also concerned about the implications and workload involved.
I guess only time will tell, and I'm in support of anything that helps OS X continue its technological advancement over rival OS's.
I guess only time will tell, and I'm in support of anything that helps OS X continue its technological advancement over rival OS's.