Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Pool details

Anyone have an idea of how this works with faster + slower drives? In a normal RAID I think you'd usually have them all the same speed at least. But say you have a few 7200rpm and some slower 5k laptop drives etc and you pool them together. Do you actually end up loosing speed? Or possibly you can't even create a pool with drives of varying speeds?
 
Are you telling us you have 500gig of data which you don't take a backup of?
Everyone should have a backup drive of at lest equal in size to their /Users directory, better still their entire drive. best their entire drive + more for incremental backups.

No, I'm telling you I don't have 500 gigs of free space FREE to copy my boot drive to. I don't really want to have to delete my backup drive just to upgrade, nor would I want to upgrade by wiping my boot drive and restoring from backup.

Not to mention that I have a huge amount of data installed from DVD's that obviously isn't backed up beyond the original install disks. It took hours to install those and I'd hate to have to do it again (or find the space on another drive to copy those over).

So is this upgrade really going to require wiping the disk? And would it be all disks or just the boot disk?
 
Heck, even using case-sensitive HFS+ is not recommended for the system drive because some apps may assume case-insensitivity and expect the files "Readme" and "README" to be the same. It's bad programming on the app developer's part, but I'm sure it's not uncommon. Macs have been case-insensitive for what, 20 years or more?

Well, to nitpick... OS X isn't completely case-insensitive. From the command line it is case-sensitive like pretty much any other Unix variant. Try using the built-in "locate" function sometime - really annoying if, like me, you come from a Linux background.

One of the first thngs I did was figure out how to build slocate on my Mac, so I can at least run "locate -i" now...
 
File Snapshots?

As I recall, Time Machine takes a snapshot of the file system once per day. ZFS saves snapshots by just storing the changed information.

But would it be possible with ZFS to just save a snapshot of a large file or set of files once every few minutes, or on request, and have it just store the changed information?


Also, I would prefer to use Time Machine with networked storage, but it seems that would be difficult unless I bought a Mini just for networking ZFS Time Machine storage, or having networked storage searchable with Spotlight. That is, if the Mini isn't discontinued as has been rumored.
 
MacFUSE is open source so they should be able to use it

Ahhh, I thought MacFUSE was licensed in such a way that it could not be incorporated into commercial products that do not provide source, without permission, but I see I am wrong! So then, I think ZFS plus MacFUSE would be a really amazing addition to Leopard's FS capabilities!
 
Rent-A-RAID

Not specifically pertaining to a ZFS update but can someone please start a business of renting a few TB drives for a week? Archive and install is nice and usually works well. But for major OS releases a clean install has always proved its value.

Too bad I have way too much data these days to backup on removable disks and no spare drives laying around. If only I could borrow a 2 TB drive for a week or so. Any takers? :rolleyes:
 
WOW. Here's something that no one has pointed out...

wikipedia said:
Pools and their associated ZFS file systems can be moved between different platform architectures, even between systems implementing different byte orders. The ZFS block pointer format allows for filesystem metadata to be stored in an endian-adaptive way; individual metadata blocks are written with the native byte order of the system writing the block. When reading, if the stored endianness doesn't match the endianness of the system, the metadata is byte-swapped in memory.

I'd think this could dramatically simplify some aspects of writing Universal applications.
 
Yes!

ZFS in Leopard! Awsome news! Is there any video, etc. of the guy leaking this info? When I go to the webcast link in the story it is of a guy in the mountains telling how to take pictures. WTF ;)
 
Anyone have an idea of how this works with faster + slower drives? In a normal RAID I think you'd usually have them all the same speed at least. But say you have a few 7200rpm and some slower 5k laptop drives etc and you pool them together. Do you actually end up loosing speed? Or possibly you can't even create a pool with drives of varying speeds?

I'd expect there would be a performance issue although the striping might mitigate this to a certain extent. I haven't read anything that suggests drive speeds are an issue. The more consistent the drives perform regarding one another the better your array would likely perform though.
 
.mac / zfs

The snapshotting features of these kind of file systems are v.cool and saving differences in large files rather than the whole file again is obviously hugely efficient .....which is great when saving over slow WAN links.....like internet.......you see where it's going................... q the new '.mac' storage on internet.........zfs enabled (i.e. save only the differences at block level of the slightly changed large file.)...... Wouldn't suprise me if time machine was linked to new .mac too.......Steve has been promising improvements......

mrfrosty
 
so how many separate drives does raid-z need to work?

i assume at least 3?

Here's the man page info:

A variation on RAID-5 that allows for better distribution of parity and eliminates the “RAID-5 write hole” (in which data and parity become inconsistent after a power loss). Data and parity is striped across all disks within a raidz group.

A raidz group can have either single- or double-parity, meaning that the raidz group can sustain one or two failures respectively without losing any data. The raidz1 vdev type specifies a single-parity raidz group and the raidz2 vdev type specifies a double-parity raidz group. The raidz vdev type is an alias for raidz1.

A raidz group with N disks of size X with P parity disks can hold approximately (N-P)*X bytes and can withstand one device failing before data integrity is compromised. The minimum number of devices in a raidz group is one more than the number of parity disks. The recommended number is between 3 and 9.

Sounds like two is possible but you might as well mirror because of the loss in space due to parity. 3 would be what I would recommend as the minimum.
 
iMac with multiple hard drives?

Anyone want to bet that the new iMac design will incorporate multiple hard drive bays? Sure would help to show off Time Machine and ZFS. ;)

and/or in the mac mini replacement / apple TV - just add more drives as your media collection grows!
 
Anyone want to bet that the new iMac design will incorporate multiple hard drive bays? Sure would help to show off Time Machine and ZFS. ;)

and/or in the mac mini replacement / apple TV - just add more drives as your media collection grows!

ZFS could be used across external drives. This would not be really great for the AppleTV since it uses USB. Although an iMac with firewire wouldn't necessarily need more drive bays since you could just chain external firewire drives to it. But more drive bays would be nice.

Also, Time Machine could work just fine with a single pool, single drive ZFS.
 
How will this new file system affect, if at all, Network Attached Storage? I have an Infrant ReadyNAS NV+ RAID system that's on my network. Would/could it be reformatted to ZFS? If so, would that affect accessing the NAS when booted into Windows via Boot Camp, let's say?
 
It seems that ZFS is one of the "secret Leopard features" we've been waiting to hear about at WWDC. How many more will leak from Apple partners?

Anybody remember when a shot of Steve with a new iMac was leaked by Time Magazine Canada?
 
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.