Welcome to the MacNN Forums.

If this is your first visit, be sure to check out the FAQ by clicking the link above. You may have to register before you can post: click the register link above to proceed. To start viewing messages, select the forum that you want to visit from the selection below.

You are here: MacNN Forums > Software - Troubleshooting and Discussion > macOS > 'Virtual' or 'logical' filesystem?

'Virtual' or 'logical' filesystem?
Thread Tools
kman42
Professional Poster
Join Date: Sep 2000
Location: San Francisco
Status: Offline
Reply With Quote
Dec 7, 2006, 02:27 PM
 
I'm not sure if this exists or if it would be advantageous, but I was helping my friend add an external drive the other day and thought of it. Right now all of the filesystems I know about start with the physical drive as the paradigm. For instance, you have an internal drive with all of the directories, including User folders, but if you add a second drive then you now have two places to put stuff one of which is outside of the User directory paradigm. While this may seem obvious and useful for external drives that you move around, it is counterintuitive for many users' setups. It seems to me that a virtual storage system would be better in some cases. You add a second drive and you just get more space on your computer. Your directory structure would be completely unchanged, you would still put stuff in your User folder, but now you would have more space. Are there filesystems like this?

I was thinking of this in the context of Time Machine. If you add an external drive it apparently asks if you want to use it for Time Machine. It would seem a small step that every time you added a drive it asked you if you wanted to add it to your main storage, Time Machine, or as a removable drive.

Dumb idea? Someone must have thought of it and we must not have it for a reason...

kman
     
krove
Mac Elite
Join Date: Jul 2000
Location: Washington, DC
Status: Offline
Reply With Quote
Dec 7, 2006, 02:55 PM
 
It's called ZFS.

A few months back, a Mac OS X engineer posted on a Sun emailing list about ZFS interest at Apple.

How did it come to this? Goodbye PowerPC. | sensory output
     
kman42  (op)
Professional Poster
Join Date: Sep 2000
Location: San Francisco
Status: Offline
Reply With Quote
Dec 7, 2006, 03:16 PM
 
Hey yeah, that's it. Cool.

From your link:
"ZFS presents a pooled storage model that completely eliminates the concept of volumes and the associated problems of partitions, provisioning, wasted bandwidth and stranded storage. Thousands of filesystems can draw from a common storage pool, each one consuming only as much space as it actually needs. The combined I/O bandwidth of all devices in the pool is available to all filesystems at all times."

I wonder if Apple will actually implement this. There seem to be lots of advantages (to my naive mind anyway).

kman
     
sushiism
Grizzled Veteran
Join Date: Jun 2002
Location: UK
Status: Offline
Reply With Quote
Dec 7, 2006, 04:28 PM
 
sounds dangerous
     
gradient
Mac Elite
Join Date: Aug 2005
Location: Vancouver, BC
Status: Offline
Reply With Quote
Dec 7, 2006, 04:39 PM
 
what would be dangerous about it?
     
Big Mac
Clinically Insane
Join Date: Oct 2000
Location: Los Angeles
Status: Offline
Reply With Quote
Dec 7, 2006, 05:03 PM
 
Perhaps because if one drive goes down, you're suddenly missing files in random places.

"The natural progress of things is for liberty to yield and government to gain ground." TJ
     
krove
Mac Elite
Join Date: Jul 2000
Location: Washington, DC
Status: Offline
Reply With Quote
Dec 7, 2006, 11:42 PM
 
This is no different than today's disks: if the disk goes down, so goes the data. If you have a single HD on your Mac, and the disk fails for whatever reason, the data might be suspect.

With two HDs, though, you can backup your data (either through software or hardware raid, or something like SuperDuper! to image one HD to a disk image on the second). Same goes for ZFS. Only, the setup is all automatic: two drives are told to create a mirrored storage pool, and you're done. ZFS corrects silent corruption between the mirrors, ZFS can create as many virtual HDs from this storage pool as you like, automatically mirroring the data. ZFS is extremely robust in terms of error protection, verifying contents of a write, silent data corruption between mirrored disks, etc. Watch some of the demos on my link from my post above.

That all said, there is no reason you can't be wreckless with ZFS like most users are today and create a storage pool that is NOT mirrored. With this setup, you can add a disk, and it adds it to the pool seamlessly without any change in filesystems you actually interact with: there is just more room! Of course, most users should have some backup strategy in place. Fortunately, Time Machine will fix that. Still, Time Machine + HFS+ (current file system used by Macs) hardly matches the features offered by ZFS.

ZFS truly is a file system of the future that Mac users should push Apple to adopt in some manner.

How did it come to this? Goodbye PowerPC. | sensory output
     
krove
Mac Elite
Join Date: Jul 2000
Location: Washington, DC
Status: Offline
Reply With Quote
Dec 7, 2006, 11:58 PM
 
Here is the original info about Mac OS X engineering interest in ZFS.

http://arstechnica.com/staff/fatbits.ars/2006/4/27/3777

How did it come to this? Goodbye PowerPC. | sensory output
     
monkeybrain
Mac Elite
Join Date: Nov 2001
Location: Dark Side of the Moon
Status: Offline
Reply With Quote
Dec 8, 2006, 03:32 AM
 
That's an interesting link. It seems like Apple has implemented some of this ZFS know-how in Leopard for Time Machine. But have they changed the file system? I hadn't heard about it.

I like one of the comments where someone uses the term 'Timewarp' to describe something exactly like Time Machine, and that was before Time Machine was announced.
     
jasong
Mac Elite
Join Date: Mar 2000
Location: Allston, MA, USA
Status: Offline
Reply With Quote
Dec 8, 2006, 07:58 AM
 
BigMac is right. Currently if a drive fails, it's pretty obvious immediately which one failed, and you have an idea of what's lost or suspect. If you have 3 drives transparently acting as one and one fails, how do you know which one? How do you know what files are suspect? What happens if a file is fragmented across two disks (can this happen in a virtual filesystem? I'm just thinking out load here).

I'm not down on the idea, I think it is a great idea and have often wished the mac worked similar to this. The way I would like to see it work is each disk is still a known, separate entity, but can act as virtual space. So I would have a user folder on Disk 2 with a layout that matches the one on Disk 1 (Documents, Pictures, Movies, etc.). Any files I put in my Movies folder on Disk 2 would appear in my Movies folder on Disk 1 (maybe with an icon overlay or some other way to indicate something is different) when mounted, and disappear (or be faded) when unmounted. That way I can put the most important files on my PowerBook and know they will always be with me even on the road, but have seamless access to them when I am at home.
-- Jason
     
kman42  (op)
Professional Poster
Join Date: Sep 2000
Location: San Francisco
Status: Offline
Reply With Quote
Dec 8, 2006, 11:09 AM
 
Originally Posted by jasong View Post
BigMac is right. Currently if a drive fails, it's pretty obvious immediately which one failed, and you have an idea of what's lost or suspect. If you have 3 drives transparently acting as one and one fails, how do you know which one? How do you know what files are suspect? What happens if a file is fragmented across two disks (can this happen in a virtual filesystem? I'm just thinking out load here).

I'm not down on the idea, I think it is a great idea and have often wished the mac worked similar to this. The way I would like to see it work is each disk is still a known, separate entity, but can act as virtual space. So I would have a user folder on Disk 2 with a layout that matches the one on Disk 1 (Documents, Pictures, Movies, etc.). Any files I put in my Movies folder on Disk 2 would appear in my Movies folder on Disk 1 (maybe with an icon overlay or some other way to indicate something is different) when mounted, and disappear (or be faded) when unmounted. That way I can put the most important files on my PowerBook and know they will always be with me even on the road, but have seamless access to them when I am at home.
I don't think the idea is to preclude a safe backup protocol. I think you still need two volumes (both virtual and possibly made up of more than one drive), where one acts as the backup for the other. This would insure duplication of data on more than one drive. It would just be easier to add an additional drive to either your main virtual volume or your backup virtual volume. If you continually add drives to only your main volume and don't setup a backup volume then you are screwed just like you are now if you only have one drive and no backup plan.

I watched part of the demo on mute before realizing there was a narrative, but it seems that was how it was designed. I could have misinterpreted it though so please correct me if I'm wrong. I'll watch it in full this weekend.

kman
     
srfdriver22
Forum Regular
Join Date: Sep 2003
Status: Offline
Reply With Quote
Dec 8, 2006, 11:46 AM
 
What happens with laptops? Say you plug in an external HD, but then unplug it to be mobile with your laptop? You'd lose those files and perhaps part of your OS! This may just be better for those non-mobile folks.
     
P
Moderator
Join Date: Apr 2000
Location: Gothenburg, Sweden
Status: Offline
Reply With Quote
Dec 8, 2006, 11:58 AM
 
ZFS implements RAID-Z, a system based on the RAID 5 concept of parity information spread over all the drives. You can lose any one drive and not lose any data. ZFS is really great, but it's not really for the desktop user - at least not yet. Too much trouble to setup.
     
kman42  (op)
Professional Poster
Join Date: Sep 2000
Location: San Francisco
Status: Offline
Reply With Quote
Dec 8, 2006, 12:31 PM
 
Originally Posted by P View Post
...but it's not really for the desktop user - at least not yet. Too much trouble to setup.
Isn't this what Apple does best? Take cool, but difficult, technology and make it useable.

kman
     
jasong
Mac Elite
Join Date: Mar 2000
Location: Allston, MA, USA
Status: Offline
Reply With Quote
Dec 8, 2006, 05:39 PM
 
I took the OP comment as this: Say you have one HD with 160GB of space. Currently if you add a second HD (say a 250GB), you have one drive of 160 and one of 250 and you have to manually place files on either drive. What he is looking for is a way for the filesystem to say you now have one HD with 410GB of space.

I don't know if this is what ZFS does, but your description doesn't make it sound like it.
-- Jason
     
kman42  (op)
Professional Poster
Join Date: Sep 2000
Location: San Francisco
Status: Offline
Reply With Quote
Dec 8, 2006, 06:52 PM
 
Originally Posted by jasong View Post
I took the OP comment as this: Say you have one HD with 160GB of space. Currently if you add a second HD (say a 250GB), you have one drive of 160 and one of 250 and you have to manually place files on either drive. What he is looking for is a way for the filesystem to say you now have one HD with 410GB of space.

I don't know if this is what ZFS does, but your description doesn't make it sound like it.
Actually, I think that is exactly what it does. You can create a 'pooled' mount consisting of multiple physical drives, but appearing as a single directory structure. You can then add more drives as you see fit.

What's not clear to me is how the mirroring works. It sounds like each bit of data is mirrored on a separate drive to protect against physical failures, but that the mirror could be on any number of different drives. That's what isn't apparent to me: what does zfs use to determine where to mirror the data? It sounds automatic and transparent, but I don't know the actual implementation.

kman
     
P
Moderator
Join Date: Apr 2000
Location: Gothenburg, Sweden
Status: Offline
Reply With Quote
Dec 11, 2006, 01:32 PM
 
Originally Posted by kman42 View Post
Isn't this what Apple does best? Take cool, but difficult, technology and make it useable.

kman
Yes. But multiple HDs are too expensive for now, especially if you need 5 or so.
     
P
Moderator
Join Date: Apr 2000
Location: Gothenburg, Sweden
Status: Offline
Reply With Quote
Dec 11, 2006, 01:51 PM
 
Originally Posted by kman42 View Post
Actually, I think that is exactly what it does. You can create a 'pooled' mount consisting of multiple physical drives, but appearing as a single directory structure. You can then add more drives as you see fit.

What's not clear to me is how the mirroring works. It sounds like each bit of data is mirrored on a separate drive to protect against physical failures, but that the mirror could be on any number of different drives. That's what isn't apparent to me: what does zfs use to determine where to mirror the data? It sounds automatic and transparent, but I don't know the actual implementation.

kman
You don't have to understand how it works, just now that it does, but let's see if I can explain it to you:

Back in the olden days of modems, you could have parity bits enabled to ease error correction. Over shaky phone lines, this helped a lot. It worked by taking ASCII letters (always 7 bits - 8-bit text uses one of several expansions of ASCII, even though they're usually also called ASCII) and adding one extra bit. This bit was set to either 1 or 0 so that there was always an even number of 1s (for instance). If the receiving computer got a byte with an odd number of 1s, it knew that the byte was bad and asked for a resend.

That's parity error detection. Error correction is a little trickier, but works the same way - only with more parity bits. Say that you want to encode every number between 0 and 15 in this way. The regular way is something like this:

0 0000
1 0001
2 0010
3 0011

etc. Now say that you add three more bits like this:

0 0000 000
1 0001 011
2 0010 101
3 0011 110

Now every number is at least 3 bits away from every other number (and I could go on to 15 and this would still be true). If any one bit is wrong - either in the first 4 figures that are the "regular" encoding or in the last three in the parity information - then that is an illegal combination. Here's the magic - I know which combination it was likely to be, because there is only one combination that is exactly one bit away.

Now store each of those 7 bits on different drives, and you have a system that will let you restore the file if any one of those drives die. RAID 5 is way more complex than that - there is much less parity information, so less waste, but also more security - but it's the same principle
     
Brass
Professional Poster
Join Date: Nov 2000
Location: Tasmania, Australia
Status: Offline
Reply With Quote
Dec 11, 2006, 05:02 PM
 
Sheesh, you don't need ZFS just to make two drives behave like one. We've had RAID that can do this for yonks. Just use a RAID utility to concatenate the second drive onto the first, and then growfs the existing file system on the fly to be able to use the increased space. Sure it's not automatic, but it's not rocket science, either.

Having said that, I've never used RAID on a Mac before, so maybe Apple have made it too hard to use (I can't imagin this), or they don't include a concatenation option?
     
mduell
Posting Junkie
Join Date: Oct 2005
Location: Houston, TX
Status: Offline
Reply With Quote
Dec 11, 2006, 07:28 PM
 
Congratulations, you've re-thought-up JBOD. It's like RAID without the safety or performance.
     
ntsc
Senior User
Join Date: Sep 2000
Location: Glasgow, Scotland UK
Status: Offline
Reply With Quote
Dec 12, 2006, 12:29 AM
 
Quote:
Originally Posted by kman42
Actually, I think that is exactly what it does. You can create a 'pooled' mount consisting of multiple physical drives, but appearing as a single directory structure. You can then add more drives as you see fit.

What's not clear to me is how the mirroring works. It sounds like each bit of data is mirrored on a separate drive to protect against physical failures, but that the mirror could be on any number of different drives. That's what isn't apparent to me: what does zfs use to determine where to mirror the data? It sounds automatic and transparent, but I don't know the actual implementation.

kman
You don't have to understand how it works, just now that it does, but let's see if I can explain it to you:

Back in the olden days of modems, you could have parity bits enabled to ease error correction. Over shaky phone lines, this helped a lot. It worked by taking ASCII letters (always 7 bits - 8-bit text uses one of several expansions of ASCII, even though they're usually also called ASCII) and adding one extra bit. This bit was set to either 1 or 0 so that there was always an even number of 1s (for instance). If the receiving computer got a byte with an odd number of 1s, it knew that the byte was bad and asked for a resend.

That's parity error detection. Error correction is a little trickier, but works the same way - only with more parity bits. Say that you want to encode every number between 0 and 15 in this way. The regular way is something like this:

0 0000
1 0001
2 0010
3 0011

etc. Now say that you add three more bits like this:

0 0000 000
1 0001 011
2 0010 101
3 0011 110

Now every number is at least 3 bits away from every other number (and I could go on to 15 and this would still be true). If any one bit is wrong - either in the first 4 figures that are the "regular" encoding or in the last three in the parity information - then that is an illegal combination. Here's the magic - I know which combination it was likely to be, because there is only one combination that is exactly one bit away.

Now store each of those 7 bits on different drives, and you have a system that will let you restore the file if any one of those drives die. RAID 5 is way more complex than that - there is much less parity information, so less waste, but also more security - but it's the same principle
Actually you've essentially described RAID 5 and raidz (the ZFS implementation) more or less. RAID 5 takes at least three disks to implement which includes a copy of half a chunk of data on one disk and the other half one the second disk. This allows parallel reads from the disks which gives the advantage of speed. In addition the third disk is used to store parity information in the way that you describe, this is used to reconstruct the data on one of the other disks should it fail. The zfs implementation available on the current version of Solaris 10 is complete but lacks some interesting features which are available in the OpenSolaris implementation. These include raidz2 which is double parity RAID (has another disk used to store error correction information), which is basically RAID 6 and it also includes the availability of hot spares. This means that you could make a RAID 5 with 4 disks where two disks are used to store data, one disk for parity/error correcting information and a fourth disk which is used to immediately reconstruct data to when one of the other disks fail. This means that you could have two disks fail before any data corruption occurs.

Mirroring literally means that the two copies of the data are stored on the mirrors. This has been the case with almost every filesystem or meta filesystem implemented.

Apple really should implement ZFS even if its just for upgrading the filesystem. It could be implemented with a pool for each physical device which would allow at transition to ZFS without a real change in semantics. Although it does ignore some of the major benefits of zfs. Interestingly zfs is so unconnected to the physical device stuff that one can create a storage pool inside a file, which is rather nice, although why one would ever want this i'm not sure...
"You can't waste a life hating people, because all they do is live their life, laughing, doing more evil."

-ALPHA ROBERTSON,whose daughter was one of four girls killed in the bombing of a Birmingham, Ala., church in 1963.
     
bloodline
Forum Regular
Join Date: Jan 2006
Location: Not far from a shop that sells Logic Pro
Status: Offline
Reply With Quote
Dec 12, 2006, 09:51 AM
 
kman, just set up the drives as a "RAID 0"...
2.8 Ghz Unibody MacBook Pro 15" - 4GB Ram - Logic Pro 8.0
2.33 Ghz C2D MacBook Pro 15" - 3GB Ram - Logic Pro 7.2
1.5 Ghz G4 PowerBook 12" - 1.25GB Ram
     
Mithras
Professional Poster
Join Date: Oct 1999
Location: :ИOITAↃOâ…ƒ
Status: Offline
Reply With Quote
Dec 12, 2006, 11:05 AM
 
Actually, RAID 0 wouldn't use all of the available space, since it requires each disk to be the same size. As mduell notes, a JBOD setup will handle virtual "spanning" of multiple physical disks. And as he also notes, it's not very good.
     
kick52
Baninated
Join Date: May 2005
Location: England
Status: Offline
Reply With Quote
Dec 12, 2006, 12:45 PM
 
that sounds pretty neat..

the only thing is the HD failing thingy... but then again, it would be pretty much the same as one drive failing..

i had my old 20gb in with my powermac for a while, but took it out as A) it was noisy! and B) i didnt like two disks on my screen, and i had all my data on the primary HD, so there was no use for it.

if ZFS came in OS X, it would be useful for a lot of people.

maybe you could select what data you wanted on each HD (like a finder info), or just have it automatic.
     
Millennium
Clinically Insane
Join Date: Nov 1999
Status: Offline
Reply With Quote
Dec 12, 2006, 02:17 PM
 
Actually, what you've described is just how Unix filesystems work. As far as the computer is concerned, there is only the one filesystem, the root of which is the root of your boot drive. When you mount other drives, you attach them to directories which are then considered mount points, but as far as interacting with the filesystem is concerned these still look like ordinary directories.

Under the hood, OSX also works this way. Every external drive you attach is mounted in a folder at /Volumes/drivename, and if you use the command-line to browse through here you won't see any difference from an ordinary folder. The old 10.0 Public Beta used to mount drives at just /drivename, which caused problems with people who had named a drive "Applications"; the /Volumes folder was created to stop this. In theory it is still possible to mount drives at other folders, though this is a difficult task.

In other words, at least from the perspective of the OS, we already have this. The Finder was coded to remove this abstraction and rebuild the more traditional Mac experience on top of it, but it sounds as though you want this abstraction removed. I'm afraid I don't know of any filesystem browser that does this at the moment.
You are in Soviet Russia. It is dark. Grue is likely to be eaten by YOU!
     
kman42  (op)
Professional Poster
Join Date: Sep 2000
Location: San Francisco
Status: Offline
Reply With Quote
Dec 16, 2006, 08:12 PM
 
     
Chuckit
Clinically Insane
Join Date: Oct 2001
Location: San Diego, CA, USA
Status: Offline
Reply With Quote
Dec 16, 2006, 08:41 PM
 
Originally Posted by kman42 View Post
Check this out!

Zettabyte sur L�opard
That page seems to be written in some kind of secret code that looks like gibberish to the ordinary observer.
Chuck
___
"Instead of either 'multi-talented' or 'multitalented' use 'bisexual'."
     
kman42  (op)
Professional Poster
Join Date: Sep 2000
Location: San Francisco
Status: Offline
Reply With Quote
Dec 16, 2006, 08:59 PM
 
By gibberish, you mean French?

I think this is the relevant part:

     
cgc
Professional Poster
Join Date: Mar 2003
Location: Down by the river
Status: Offline
Reply With Quote
Dec 17, 2006, 02:12 AM
 
Originally Posted by jasong View Post
BigMac is right. Currently if a drive fails, it's pretty obvious immediately which one failed, and you have an idea of what's lost or suspect. If you have 3 drives transparently acting as one and one fails, how do you know which one? How do you know what files are suspect? What happens if a file is fragmented across two disks (can this happen in a virtual filesystem? I'm just thinking out load here).
...
Any drive supporting S.M.A.R.T., which should include any fairly recent models, will let you know it's beginning to fail or has failed. S.M.A.R.T. detected my primary HDD was failing and allowed me to do a complete transfer (at a ludicrously slow rate due to the impending failure) prior to it being completely useless.

One problem though is that I found the S.M.A.R.T. alert by going into Disk Utility. It'd be nice if Apple would pop up a window (ala Growl) telling us our hard drive is failing.
( Last edited by cgc; Dec 17, 2006 at 02:25 AM. Reason: Adding missing detail)
     
mac1896
Forum Regular
Join Date: Dec 2005
Location: Florida
Status: Offline
Reply With Quote
Dec 17, 2006, 02:50 AM
 
Originally Posted by cgc View Post
Any drive supporting S.M.A.R.T., which should include any fairly recent models, will let you know it's beginning to fail or has failed. S.M.A.R.T. detected my primary HDD was failing and allowed me to do a complete transfer (at a ludicrously slow rate due to the impending failure) prior to it being completely useless.

One problem though is that I found the S.M.A.R.T. alert by going into Disk Utility. It'd be nice if Apple would pop up a window (ala Growl) telling us our hard drive is failing.
If you don't mind using a 3rd-party app, that places a status icon in your menubar, SMARTReporter is great.

Got mine set to run at startup, but if you choose, you can run it only occasionaly (which, to me, is sensless if you want to monitor your drive's status).

Free Software Development for Mac OS X

.......and it's FREE



I now return you to the discussion already in progress.............
( Last edited by mac1896; Dec 17, 2006 at 02:54 AM. Reason: forgot link)
Gee, I hope they're friendly..........
     
Gee4orce
Professional Poster
Join Date: Dec 2000
Location: Staffs, UK
Status: Offline
Reply With Quote
Dec 17, 2006, 12:11 PM
 


Overall ZFS offers the following key advantages:-

Pooled storage - No requirement for a volume manager when extra volumes added, the volume is simply added to a pool creating a vdev (virtual device), a collection of vdevs makes up a zpool, which in essence is the storage available to the file system.

Snapshots - Read-only point in time of the file system

Clones - write-able copy of a snapshot

RAID-Z - Makes use of copy-on-write; rather than overwriting old data with new data, it writes new data to a new location and then overwrites the pointer to the old data

Detects and then corrects data corruption

Incredibly fast due to intelligent pre-fetching, and dynamic striping
More ZFS info:

ZFS has been subjected to over a million forced, violent crashes without losing data integrity or leaking a single block.
Also ZFS can address more storage than it's physically possible to have given the number of atoms available in the entire Earth ! more realistically, you could pool all the hard drives in the world into a single giant storage pool and still uniquely address each and every bit and byte. Pretty impressive.
( Last edited by Gee4orce; Dec 17, 2006 at 01:09 PM. )
     
mduell
Posting Junkie
Join Date: Oct 2005
Location: Houston, TX
Status: Offline
Reply With Quote
Dec 17, 2006, 03:25 PM
 
Originally Posted by Gee4orce View Post
Also ZFS can address more storage than it's physically possible to have given the number of atoms available in the entire Earth ! more realistically, you could pool all the hard drives in the world into a single giant storage pool and still uniquely address each and every bit and byte. Pretty impressive.
ZFS supports files and volumes up to 16 EiB... the same as NTFS and HFS+. UFS2 supports about a 100000x more storage than ZFS.
     
P
Moderator
Join Date: Apr 2000
Location: Gothenburg, Sweden
Status: Offline
Reply With Quote
Dec 17, 2006, 04:34 PM
 
Originally Posted by mduell View Post
ZFS supports files and volumes up to 16 EiB... the same as NTFS and HFS+. UFS2 supports about a 100000x more storage than ZFS.
So we should really be calling it 0.016ZFS, then? (Z = Zettabyte = 1000 Exabytes)
     
Gee4orce
Professional Poster
Join Date: Dec 2000
Location: Staffs, UK
Status: Offline
Reply With Quote
Dec 17, 2006, 05:02 PM
 
I'm don't know anything about UFS2, but saying it supports 100000x more storage than ZFS is like saying you'd be 100000x more dead if you were hit by a train than by a car ! Like I said, it'd be impossible to fill a ZFS filesystem without resorting to Star Trek like sci-fi.

The maximum size of a single file is in the exabyte range. The theoretical total size of the storage pool exceeds a zetabyte - hence the name.

Sounds like it's going to be an exciting addition to the Mac OS X - I wonder if it'll replace HFS+ - maybe in Mac OS XI ?
     
Ganesha
Senior User
Join Date: Jul 2002
Location: Arizona Wasteland
Status: Offline
Reply With Quote
Dec 17, 2006, 05:17 PM
 
But the theoretical limit to the size of the ZFS pool is 3 x 10^18 Zettabytes. Out of this pool can be carved many (up to 2^64) filesystems with a limit of 16 EB.
     
Brass
Professional Poster
Join Date: Nov 2000
Location: Tasmania, Australia
Status: Offline
Reply With Quote
Dec 18, 2006, 07:40 PM
 
Originally Posted by kick52 View Post
that sounds pretty neat..

the only thing is the HD failing thingy... but then again, it would be pretty much the same as one drive failing..
Not quite the same... with twice the number of disks, you are twice as likely to have the system fail catastrophically. Which is why RAID 1 or 5 is better (but wastes more disk space). RAID 0 and/or JBOD provides no redundancy, and is LESS safe than a single disk.
     
   
Thread Tools
 
Forum Links
Forum Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts
BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Top
Privacy Policy
All times are GMT -4. The time now is 04:34 AM.
All contents of these forums © 1995-2017 MacNN. All rights reserved.
Branding + Design: www.gesamtbild.com
vBulletin v.3.8.8 © 2000-2017, Jelsoft Enterprises Ltd.,