We're not going to have to wait for ZFS folks. Remember the WWDC beta is going to be about a full month later than the last Beta in early March. ZFS was workable two betas ago.
Sun is booting OpenSolaris on ZFS and Thumper uses it already production hardware.
hmurch is right. ZFS has had a lot of time to mature. And with Leopard's release in 5 months, all the remaining problems (if there still are some) will be solved. This FS has been around since 2004...has been up and running on OS X near the end of December/early January and will have had over 10 months of work put into making it work well. If that's not enough time, how long IS enough? 5 years? 10?
hmurch is right. ZFS has had a lot of time to mature. And with Leopard's release in 5 months, all the remaining problems (if there still are some) will be solved. This FS has been around since 2004...has been up and running on OS X near the end of December/early January and will have had over 10 months of work put into making it work well. If that's not enough time, how long IS enough? 5 years? 10?
It's interesting that even people in the position of Schwartz can get excited over the prospect of something that they consider to be significant, and blurt out something, that, after consideration, might have been a bit more than it should have been.
I haven't programmed for a while, but, I can see the issues arising with this being THE default FS, right now.
That doesn't mean that Apple hasn't solved them.
Hopefully we will find out tomorrow.
I'm going to fly to California early tomorrow morning, so that I can hear Job's presentation three hours earlier than I would find out here in NYC, because of the time Zone difference.
Because when Schwartz spilled the beans, Jobs called him and screamed at him for doing so. Schwartz then had to backtrack.
It's totally obvious that ZFS as the default filesystem in Leoparad is one of the top secret features that Jobs wanted to shock the world with.
I doubt this for the reason that consumers don't know or give a rip about filesystems. Jobs could blather on and on but consumers would never get excited because bragging about filesystems is like bragging about kernels. It's boring stuff.
I doubt ZFS would be used for boot volumes for Leopard, but there are two options still...
The first is just using ZFS by default for Time Machine. This is pretty obvious and has been discussed to death.
The second is using a flash boot volume w/o ZFS and then using ZFS on the main disk storage including user home directories, Apps, etc. This way we get the fast booting and other goodies from the flash upgrade and get the ZFS goodness w/o addressing the boot issue.
Bugs aside, it's entirely conceivable that Leopard got pushed back for ZFS. ZFS, while not a marketable feature in itself, is probably one of the most important Leopard feature. It's the backbone for a ton of new features that *are* consumer features.
Time Machine was a cool idea, but when I first heard how it was actually implemented on HFS+, I thought "This is some kind of hack job, it's not gonna work very well." And it wouldn't have worked very well...it would have been portable in the sense that it would have worked on most file systems but the performance and the efficiency would have been lost by the very nature of how it handled snapshots.
The kernel isn't a marketable feature...but a crash-free OS *is*. So is a solid backup system that's basically hands free. So is a storage system that is as easy to manage as adding RAM to a computer.
Sure, one excuse for the delay might have been the ridiculous amount of bugs in Leopard...but that's never stopped Apple from shipping x.0 software. ZFS is one of those more believable excuses. It has to be there in x.0 and bugs with it would have been a show-stopper for everyone...not just a hand full of people.
Bugs aside, it's entirely conceivable that Leopard got pushed back for ZFS. ZFS, while not a marketable feature in itself, is probably one of the most important Leopard feature. It's the backbone for a ton of new features that *are* consumer features.
Time Machine was a cool idea, but when I first heard how it was actually implemented on HFS+, I thought "This is some kind of hack job, it's not gonna work very well." And it wouldn't have worked very well...it would have been portable in the sense that it would have worked on most file systems but the performance and the efficiency would have been lost by the very nature of how it handled snapshots.
The kernel isn't a marketable feature...but a crash-free OS *is*. So is a solid backup system that's basically hands free. So is a storage system that is as easy to manage as adding RAM to a computer.
Sure, one excuse for the delay might have been the ridiculous amount of bugs in Leopard...but that's never stopped Apple from shipping x.0 software. ZFS is one of those more believable excuses. It has to be there in x.0 and bugs with it would have been a show-stopper for everyone...not just a hand full of people.
What else would benefit (feature wise) from ZFS file system? Could spotlight be improved?
What else would benefit (feature wise) from ZFS file system? Could spotlight be improved?
Anything that accesses disks could be improved especially applications that cache a lot of data like Photoshop. As they say:
"The architecture in ZFS removes just about all constraints on I/O order, which allows us to run the disks much closer to full I/O capacity than current filesystems/volume managers."
One thing I was worried about was fragmentation but according to their FAQ, the allocation algorithms handle this. So maybe this would be an end to the Bootcamp partitioning problems.
Anything that accesses disks could be improved especially applications that cache a lot of data like Photoshop. As they say:
"The architecture in ZFS removes just about all constraints on I/O order, which allows us to run the disks much closer to full I/O capacity than current filesystems/volume managers."
One thing I was worried about was fragmentation but according to their FAQ, the allocation algorithms handle this. So maybe this would be an end to the Bootcamp partitioning problems.
A: It won't be in the first ZFS release; look for it in a Solaris 10 OS update.
Clearly, OS X won't be using ZFS as the boot/default filesystem. Hell, Solaris doesn't even have it bootable.
This Sun document on ZFS installation has some tidbits that I can see people inferring from [Time Machine being one of them], but there is obvious a lot of information lacking:
Sun specifically discusses UFS to ZFS migration and nothing else.
Since HFS+ is the preferred format for OS X then Apple would have had to do all the work for Sun as well as do joint collaboration to expose from both companies the innards of both filesystems to make a HFS+ -> ZFS conversion possible.
I can see ZFS support being included into OS X Server and some extensive GUI Tools to help administer ZFS Pools, but I don't see OS X Leopard Client moving to ZFS as the default filesystem.
Look how old that FAQ is, it was released when ZFS was first announced. Solaris 10 wasn't even out by then. Seriously, nothing on those pages is up to date. ZFS has been bootable for a few weeks at least, and the issues for getting OS X booting off of it are different than for Solaris. Apple has been working on this for a while now and could have had booting working for quite some time now.
Bugs aside, it's entirely conceivable that Leopard got pushed back for ZFS. ZFS, while not a marketable feature in itself, is probably one of the most important Leopard feature. It's the backbone for a ton of new features that *are* consumer features.
Time Machine was a cool idea, but when I first heard how it was actually implemented on HFS+, I thought "This is some kind of hack job, it's not gonna work very well." And it wouldn't have worked very well...it would have been portable in the sense that it would have worked on most file systems but the performance and the efficiency would have been lost by the very nature of how it handled snapshots.
The kernel isn't a marketable feature...but a crash-free OS *is*. So is a solid backup system that's basically hands free. So is a storage system that is as easy to manage as adding RAM to a computer.
Sure, one excuse for the delay might have been the ridiculous amount of bugs in Leopard...but that's never stopped Apple from shipping x.0 software. ZFS is one of those more believable excuses. It has to be there in x.0 and bugs with it would have been a show-stopper for everyone...not just a hand full of people.
Those are really well-tought comments, kim kap sol, but I still have doubts. I don't believe Apple would risk such a move without having clear benefits for the consumers, capable to counter-balance the possible shortcomings. A crash-free OS would probably be such a feature, but what "a crash-free OS" is supposed to mean here? Alright, some hours yet and we will know more.
Look how old that FAQ is, it was released when ZFS was first announced. Solaris 10 wasn't even out by then. Seriously, nothing on those pages is up to date. ZFS has been bootable for a few weeks at least, and the issues for getting OS X booting off of it are different than for Solaris. Apple has been working on this for a while now and could have had booting working for quite some time now.
Yep, it says on wikipedia:
"ZFS is currently not available as a root filesystem on Solaris 10, since there is no ZFS boot support. The ZFS Boot project recently successfully added boot support to the OpenSolaris project, and is available in recent builds of Solaris Nevada.[10][11] ZFS boot is currently (20070208) planned for a Solaris 10 update in late 2007."
Late 2007 sounds familiar so it will definitely be ready by the time leopard is shipping if as Thinine says it isn't already. I think Apple will have it ready for the beta release and developers will need to test it out.
Leopard was not feature frozen officially yet. So it's not just lack of resources for squashing the bugs because of the engineers moved to iPhone.
Unlike Microsoft guys, Job's does not like to announce a feature until he is absolutely sure it will be implemented.
Let's hypothesize: when Jobs announced Leopard delay he wanted ZFS but there were problems identified, and he was not sure they will be resolved. Depending of the progress, he will or will not announce full ZFS support in 10.5.0. The hints from Sun indicate that ZFS will be supported, but whether it will be the default FS is still questionable. We will see in few hours. The point is, no matter on outcome, Apple is very serious about ZFS.
If Apple is almost ready with ZFS, but not 100% ready, I don't believe it will wait another 2 years for 10.6. Jobs may announce optional/developer/experimental support or some other marketing trick to make room for it in a 10.5.x release.
Sorry to be slow, I'm still unclear - will ZFS even be an option in Leopard? I wasn't aware of it being mentioned today ...
There was no public word on that. I am afraid only developers which assisted the conference will know (the only ones they will receive the Leopard beta).
Comments
We're not going to have to wait for ZFS folks. Remember the WWDC beta is going to be about a full month later than the last Beta in early March. ZFS was workable two betas ago.
Sun is booting OpenSolaris on ZFS and Thumper uses it already production hardware.
case insensitivity has been "fast tracked"
http://www.opensolaris.org/os/commun...C4644C98986C1F
With 4 months to sweeten the Leopard Beta before October I have no doubts that ZFS will be ready in October. Simply give users the choice.
1. Install HFS+ (Recommended)
2. Install ZFS (read help file about concerns)
I'd rather run Time Machine on ZFS myself.
I agree. I've also said that Apple would offer it as an option, the way they offer the UNIX file system, but recommend that most not use it.
Schwartz would not make such a mistake.
hmurch is right. ZFS has had a lot of time to mature. And with Leopard's release in 5 months, all the remaining problems (if there still are some) will be solved. This FS has been around since 2004...has been up and running on OS X near the end of December/early January and will have had over 10 months of work put into making it work well. If that's not enough time, how long IS enough? 5 years? 10?
If that's not enough time, how long IS enough? 5 years? 10?
I don't know but I contacted Microsoft to find out.
But why would Schwartz be mistaken (or lying)?
Because when Schwartz spilled the beans, Jobs called him and screamed at him for doing so. Schwartz then had to backtrack.
It's totally obvious that ZFS as the default filesystem in Leoparad is one of the top secret features that Jobs wanted to shock the world with.
But why would Schwartz be mistaken (or lying)?
Schwartz would not make such a mistake.
hmurch is right. ZFS has had a lot of time to mature. And with Leopard's release in 5 months, all the remaining problems (if there still are some) will be solved. This FS has been around since 2004...has been up and running on OS X near the end of December/early January and will have had over 10 months of work put into making it work well. If that's not enough time, how long IS enough? 5 years? 10?
It's interesting that even people in the position of Schwartz can get excited over the prospect of something that they consider to be significant, and blurt out something, that, after consideration, might have been a bit more than it should have been.
I haven't programmed for a while, but, I can see the issues arising with this being THE default FS, right now.
That doesn't mean that Apple hasn't solved them.
Hopefully we will find out tomorrow.
I'm going to fly to California early tomorrow morning, so that I can hear Job's presentation three hours earlier than I would find out here in NYC, because of the time Zone difference.
Just kidding.
Because when Schwartz spilled the beans, Jobs called him and screamed at him for doing so. Schwartz then had to backtrack.
It's totally obvious that ZFS as the default filesystem in Leoparad is one of the top secret features that Jobs wanted to shock the world with.
I doubt this for the reason that consumers don't know or give a rip about filesystems. Jobs could blather on and on but consumers would never get excited because bragging about filesystems is like bragging about kernels. It's boring stuff.
Storagemojo talks about Apple's patent for "touchless fs conversion
Hmmmmmmmmmmmmmmm
The first is just using ZFS by default for Time Machine. This is pretty obvious and has been discussed to death.
The second is using a flash boot volume w/o ZFS and then using ZFS on the main disk storage including user home directories, Apps, etc. This way we get the fast booting and other goodies from the flash upgrade and get the ZFS goodness w/o addressing the boot issue.
Time Machine was a cool idea, but when I first heard how it was actually implemented on HFS+, I thought "This is some kind of hack job, it's not gonna work very well." And it wouldn't have worked very well...it would have been portable in the sense that it would have worked on most file systems but the performance and the efficiency would have been lost by the very nature of how it handled snapshots.
The kernel isn't a marketable feature...but a crash-free OS *is*. So is a solid backup system that's basically hands free. So is a storage system that is as easy to manage as adding RAM to a computer.
Sure, one excuse for the delay might have been the ridiculous amount of bugs in Leopard...but that's never stopped Apple from shipping x.0 software. ZFS is one of those more believable excuses. It has to be there in x.0 and bugs with it would have been a show-stopper for everyone...not just a hand full of people.
Bugs aside, it's entirely conceivable that Leopard got pushed back for ZFS. ZFS, while not a marketable feature in itself, is probably one of the most important Leopard feature. It's the backbone for a ton of new features that *are* consumer features.
Time Machine was a cool idea, but when I first heard how it was actually implemented on HFS+, I thought "This is some kind of hack job, it's not gonna work very well." And it wouldn't have worked very well...it would have been portable in the sense that it would have worked on most file systems but the performance and the efficiency would have been lost by the very nature of how it handled snapshots.
The kernel isn't a marketable feature...but a crash-free OS *is*. So is a solid backup system that's basically hands free. So is a storage system that is as easy to manage as adding RAM to a computer.
Sure, one excuse for the delay might have been the ridiculous amount of bugs in Leopard...but that's never stopped Apple from shipping x.0 software. ZFS is one of those more believable excuses. It has to be there in x.0 and bugs with it would have been a show-stopper for everyone...not just a hand full of people.
What else would benefit (feature wise) from ZFS file system? Could spotlight be improved?
The kernel isn't a marketable feature...
Makes me wonder if Apple will finally dump the Mach kernel.
A MacRumors thread lambasting the newly released iPod from 23-OCT-2001...
"It's now at the online Apple Store!
$400 for an Mp3 Player!
I'd call it the Cube 2.0 as it wont sell, and be killed off in a short time...and it's not really functional.
Uuhh Steve, can I have a PDA now?"
I got a kick out it, perhaps you will too.
What else would benefit (feature wise) from ZFS file system? Could spotlight be improved?
Anything that accesses disks could be improved especially applications that cache a lot of data like Photoshop. As they say:
"The architecture in ZFS removes just about all constraints on I/O order, which allows us to run the disks much closer to full I/O capacity than current filesystems/volume managers."
One thing I was worried about was fragmentation but according to their FAQ, the allocation algorithms handle this. So maybe this would be an end to the Bootcamp partitioning problems.
http://www.sun.com/emrkt/campaign_do..._zfs_perf.html
Anything that accesses disks could be improved especially applications that cache a lot of data like Photoshop. As they say:
"The architecture in ZFS removes just about all constraints on I/O order, which allows us to run the disks much closer to full I/O capacity than current filesystems/volume managers."
One thing I was worried about was fragmentation but according to their FAQ, the allocation algorithms handle this. So maybe this would be an end to the Bootcamp partitioning problems.
http://www.sun.com/emrkt/campaign_do..._zfs_perf.html
Excerpt from Sun:
Q: Is ZFS bootable, and if not, when will it be?
A: It won't be in the first ZFS release; look for it in a Solaris 10 OS update.
Clearly, OS X won't be using ZFS as the boot/default filesystem. Hell, Solaris doesn't even have it bootable.
This Sun document on ZFS installation has some tidbits that I can see people inferring from [Time Machine being one of them], but there is obvious a lot of information lacking:
http://www.sun.com/emrkt/campaign_do...s_install.html
Sun specifically discusses UFS to ZFS migration and nothing else.
Since HFS+ is the preferred format for OS X then Apple would have had to do all the work for Sun as well as do joint collaboration to expose from both companies the innards of both filesystems to make a HFS+ -> ZFS conversion possible.
I can see ZFS support being included into OS X Server and some extensive GUI Tools to help administer ZFS Pools, but I don't see OS X Leopard Client moving to ZFS as the default filesystem.
Bugs aside, it's entirely conceivable that Leopard got pushed back for ZFS. ZFS, while not a marketable feature in itself, is probably one of the most important Leopard feature. It's the backbone for a ton of new features that *are* consumer features.
Time Machine was a cool idea, but when I first heard how it was actually implemented on HFS+, I thought "This is some kind of hack job, it's not gonna work very well." And it wouldn't have worked very well...it would have been portable in the sense that it would have worked on most file systems but the performance and the efficiency would have been lost by the very nature of how it handled snapshots.
The kernel isn't a marketable feature...but a crash-free OS *is*. So is a solid backup system that's basically hands free. So is a storage system that is as easy to manage as adding RAM to a computer.
Sure, one excuse for the delay might have been the ridiculous amount of bugs in Leopard...but that's never stopped Apple from shipping x.0 software. ZFS is one of those more believable excuses. It has to be there in x.0 and bugs with it would have been a show-stopper for everyone...not just a hand full of people.
Those are really well-tought comments, kim kap sol, but I still have doubts. I don't believe Apple would risk such a move without having clear benefits for the consumers, capable to counter-balance the possible shortcomings. A crash-free OS would probably be such a feature, but what "a crash-free OS" is supposed to mean here? Alright, some hours yet and we will know more.
Look how old that FAQ is, it was released when ZFS was first announced. Solaris 10 wasn't even out by then. Seriously, nothing on those pages is up to date. ZFS has been bootable for a few weeks at least, and the issues for getting OS X booting off of it are different than for Solaris. Apple has been working on this for a while now and could have had booting working for quite some time now.
Yep, it says on wikipedia:
"ZFS is currently not available as a root filesystem on Solaris 10, since there is no ZFS boot support. The ZFS Boot project recently successfully added boot support to the OpenSolaris project, and is available in recent builds of Solaris Nevada.[10][11] ZFS boot is currently (20070208) planned for a Solaris 10 update in late 2007."
Late 2007 sounds familiar so it will definitely be ready by the time leopard is shipping if as Thinine says it isn't already. I think Apple will have it ready for the beta release and developers will need to test it out.
Unlike Microsoft guys, Job's does not like to announce a feature until he is absolutely sure it will be implemented.
Let's hypothesize: when Jobs announced Leopard delay he wanted ZFS but there were problems identified, and he was not sure they will be resolved. Depending of the progress, he will or will not announce full ZFS support in 10.5.0. The hints from Sun indicate that ZFS will be supported, but whether it will be the default FS is still questionable. We will see in few hours. The point is, no matter on outcome, Apple is very serious about ZFS.
If Apple is almost ready with ZFS, but not 100% ready, I don't believe it will wait another 2 years for 10.6. Jobs may announce optional/developer/experimental support or some other marketing trick to make room for it in a 10.5.x release.
Sorry to be slow, I'm still unclear - will ZFS even be an option in Leopard? I wasn't aware of it being mentioned today ...
There was no public word on that. I am afraid only developers which assisted the conference will know (the only ones they will receive the Leopard beta).