How are you pooling data? No, ZFS is really pooling data and pooling your real hardware resources so you stop even having to think about them.<br><br>Say you have a box in front of you that has four bays for hard drives and they are 1TB, 1TB, 200GB, empty. And you have a server down the hall that is 100TB. <br>From your computer that 102.2TB of room is just one pool. You save, you work and you never think, "I have to save this to that drive so I can have a copy that is archived. Because you have already set it up so that all your files are redundant to the server.<br><br>Now here comes the scary part. You've had that set up for months now and your computer tells you that you really should put another TB into the empty Bay. You do and it joins the ZFS pool. Live, no restart, no setting up. <br><br>What ZFS is doing is making the pool any size you want, it error corrects so bad sectors always have error correction and drives have redundancy. Your computer says bay 2 TB is going bad and should be replaced. Just pull it out, throw it away and replace, maybe with a 5TB drive since they are cheaper. It all goes into the pool.<br><br>The concept of hard drive is gone. You may have artificial fences put up "my stuff", "my old stuff", "my backup stuff" but all of those things reside redundantly all over the place.<br><br>And with Spotlight 4.0 the idea of folders, directories, filing, all go out the window too. Why put a file in some arcane sub directory as if you still had steel file cabinets? What a waste of time. <br><br>The only problem with ZFS is the nature of man. A hoarder who won't throw stuff out unless necessary. So fouling terrabytes with useless junk and 700 copies of that movie of the dog skate boarding will fill your pool like oak leaves in the fall.<br><br>But PoolBoy 6.0 to the rescue.<br><br>
Registered: 12/06/07
Posts: 199
Loc: Carroll County, MD, USA
<blockquote><font size=1>In reply to:</font><hr><p>PoolBoy 6.0 to the rescue.<p><hr></blockquote><p>I bet that got the MacBabes' attention <br><br>Who serves his fellows, is of his fellows greatest - Chingachgook (by legend) <br><br>Scout
_________________________ Who serves his fellows, is of his fellows greatest - Chingachgook (by legend) Scout
#366420 - 06/10/0812:44 PMRe: ZFS coming to Snow Leopard Server
[Re: polymerase]
SteveS
Where's the cache?
Registered: 01/15/08
Posts: 106
"The concept of hard drive is gone. You may have artificial fences put up "my stuff", "my old stuff", "my backup stuff" but all of those things reside redundantly all over the place."<br><br>Artificial fences with redundancy aren't necessarily a bad thing. What happens when one of your drives goes? You now have a broken pool. <br><br>Anyway, all of this is good for a server environment. I'm not sure how useful most of this is for the home user. ZFS looks to be a great file system, but it's not as if the needs of Mac users have outgrown HFS+ by any means. <br><br>That said, I think some of us are jumping to conclusions with regard to how ZFS will be used. Just because OS X Server will finally support read/write access to ZFS, doesn't mean ZFS will be the default file system. I'd be more than willing to bet otherwise. I expect the transition to ZFS to be more gradual. Having ZFS as the default file system in 10.7 would be aggressive.<br><br>
I like artificial fences. Creating them they way you want instead of the way the hardware wants or is allocated is the slick premise of ZFS. And I am all for redundancy. I like about five fold redundancy on some documents and two fold on most everything else down to no redundancy for things that are on the net somewhere.<br><br>Drives don't usually go, they go one by one. I thought the big item for ZFS was smart redundancy run by the ZFS system that can recover with a drive failure. That's where it all gets sketchy.<br><br>And you're right, I don't think I will need ZFS in my house for another few years if at all. Depends on how my personal data ends up. The large items, music, photos, videos, might for the most part be offsite anyway. The rest is really not big enough to interest ZFS. <br><br>
#366422 - 06/10/0803:20 PMRe: ZFS coming to Snow Leopard Server
[Re: polymerase]
SteveS
Where's the cache?
Registered: 01/15/08
Posts: 106
"Drives don't usually go, they go one by one."<br><br>Agreed. But, when you have multiple drives creating one pool, just having one drive go will give you a broken pool. <br><br>"I thought the big item for ZFS was smart redundancy run by the ZFS system that can recover with a drive failure."<br><br>Yeah, ZFS has a cool feature of being able to report and repair inconsistencies, etc. assuming you have a RAID configuration setup. There is no magic here. Again, this is great for a server environment, but it may actually be more troublesome for end users that create storage pools of multiple disks without redundancy.<br><br>Here's a link you might want to look at. http://www.solarisinternals.com/wiki/index.php/ZFS_Best_Practices_Guide<br><br>Specifically:<br>"Additional Cautions for Storage Pools<br>Review the following cautions before building your ZFS storage pool:<br>A pool created with a single slice has no redundancy and is at risk for data loss.<br>A pool created with multiple slices but no redundancy is also at risk for data loss. A pool created with multiple slices across disks is harder to manage than a pool created with whole disks.<br>A pool created with whole disks but no redundancy is at risk for data loss. In addition, a pool that is not created with ZFS redundancy (RAIDZ or mirror) will only be able to report data inconsistencies. It will not be able to repair data inconsistencies. Finally, a pool created without ZFS redundancy is harder to manage because you cannot replace or detach disks in a non-redundant ZFS configuration.<br>A pool cannot be shared across systems. ZFS is not a cluster file system."<br><br><br><br><br>
"Why put a file in some arcane sub directory as if you still had steel file cabinets?"<br><br>'Cause the method works extremely well for those of us who are organized. Gawd forbid I should just toss files into one big can! I like my structured "filing cabinet".<br><br>- a.k.a. Mississauga -