At 8:44 AM -0600 1/5/2009, D Stubbs wrote:
>I am trying to get my head around this concept - that a partition is 
>not a partition. Tried googling
>
>But I am no further in understanding what is actually happening.
>I suppose it is some kind of 'virtual directory'?
>What makes no sense yet to me is how, say partition #3 could go bad, 
>and not the others - if it they all bunched together.

Ok.  From the top... simplified (ie, I'm leaving some details out)...

Physically... A hard drive is made of platters, stacked together, 
with space between them.  The platter is "divided" into concentric 
rings called tracks.  Each track is divided into sectors.  Depending 
on how your drive is formatted, the sectors are 512 bytes or 1024 or 
2048, etc.  Taken together, vertically, a cylinder is a column of 
tracks.  The drive's read/write heads are mounted on arms (fingers) 
that reach in to a cylinder and read or write - as much as the whole 
track at once.  Need a different sector?  Move that arm to a new 
position (seek), wait for the platter to spin around until the sector 
you need is near the read/write head (latency), then read or write.

But how do you access all that?  Between the HD mechanism, its 
on-board controller, the controller in your computer, and the 
operating system... those cylinder/track/sector designations are 
converted into Logical Block Numbers (LBN), 1 thru n.

That's still a bit unmanagable, so we divide things further into more 
managable chunks, called Partitions.

A partition is physically, more or less, a group of cylinders.
(it's not always a clean n cylinders as the user can pick any size).

When we "initialize" a drive, we're actually laying down the 
Partition Map at the beginning of the drive.  The first partition 
holds the Partition Map.  The next few hold special data - things 
like the disk's driver, parameter files, etc.  The last partitions 
are the ones that you-the-user see as mountable "volumes".  Within 
those "volumes" we lay down a File System.  Within that file system, 
you put your files.

Here is the partition map from my HD...  (from the command 'diskutil list')

    #:                   type name               size      identifier
    0: Apple_partition_scheme            *74.5 GB  disk0
    1:    Apple_partition_map               31.5 KB   disk0s1
    2:         Apple_Driver43                 28.0 KB   disk0s2
    3:         Apple_Driver43                  28.0 KB   disk0s3
    4:       Apple_Driver_ATA                28.0 KB   disk0s4
    5:       Apple_Driver_ATA                28.0 KB   disk0s5
    6:         Apple_FWDriver                 256.0 KB  disk0s6
    7:     Apple_Driver_IOKit                 256.0 KB  disk0s7
    8:          Apple_Patches                  256.0 KB  disk0s8
    9:              Apple_HFS Stuff            9.9 GB    disk0s10
   10:              Apple_HFS MacHD       64.4 GB   disk0s12

As you can see, I have two mountable volumes on this drive, one 
called Stuff and one called MacHD.

Because partitions are maintained as separately managed "things" by 
the Operating System, it is rare for any disk write to spaz out and 
scribble beyond a partition boundary.  That's why you can destroy the 
file system in one volume without hurting the one in another.  Of 
course, if there's a hardware failure, all bets are off.

>If it is a stack of discs - and if something failed does it stand to 
>reason that it would fail one disc at at time? - so one's best 
>security is to perhaps partition exactly to each disc?

Ok.  Failures...  Hard and Soft.

A hard failure is when something in the Hard Drive physically goes bad.

Blocks (sectors) go bad now and then.  Usually it's because they got 
a bit of oxidation on them.  When the controller notices (ie, you try 
to read the block) it tries to correct the data with whatever 
checksums or CRC information it has, then it maps-out that block and 
puts your data in a different location.  When a block goes bad, often 
it's not just one block.  Usually, the bad block "creep" is in the 
adjoining blocks, in the same track.  Sometimes the block on 
adjoining tracks fail too, but that's pretty rare.  (the inter-track 
gap is gigantic compared to the inter-sector gap).

Then there's a head crash... that's when one of those read/write 
heads accidentally touches the platter.  The head is damaged, the 
platter is damaged.  Worse - tiny bits of material (schrapnel) flies 
around and starts destroying other things.  ...Great high school 
physics problem, figuring the impact velocity of a bit of oxide flung 
against a platter spinning at 7200rpm...

A soft failure is when bad data is written to the drive.  This can be 
outright bad data (spazz), or omission of data (the app or OS crashes 
before it finishes sending data to the drive).  If the error is 
within the structure of a file system or the partition map itself, 
then a tool such as Disk Utility is used to "repair" it.  If the 
error is within a file's data, then it's up to the user to restore 
that file from backup.


Now... Let's put all this in context.

Back in the old days, we often recommended using lots of partitions 
to separate your data from the operating system because the classic 
Mac OS had no deep protection mechanisms.  If something spazzed out, 
it destroyed data.  The idea was to "limit" the destruction.  But 
those days are over.  With HFS+, and now HFS+ Journaled, as used by 
our modern Mac OS X... spazzing just doesn't happen.

So the ONLY reason to make extra volumes out of your disk is if YOU 
the user want that logical separation.

Leaving things as one large volume has its advantages -- a bigger 
pool of space to use for whatever.

We recommend partitioning large external disks mostly to encourage 
people to maintain their backups.  Make the first volume the same 
size as your internal drive  (well, at least big enough to hold all 
the data on that drive).  And let the 2nd volume be the rest.  Use a 
tool such as Carbon Copy Cloner or SuperDuper to make your backup of 
your internal drive to that first volume on the external....

HTH,
- Dan.
-- 
- Psychoceramic Emeritus; South Jersey, USA, Earth

--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed Low End Mac's G3-5 List, a 
group for those using G3, G4, and G5 desktop Macs - with a particular focus on 
Power Macs.
The list FAQ is at http://lowendmac.com/lists/g-list.shtml and our netiquette 
guide is at http://www.lowendmac.com/lists/netiquette.shtml
To post to this group, send email to g3-5-list@googlegroups.com
To unsubscribe from this group, send email to 
g3-5-list-unsubscr...@googlegroups.com
For more options, visit this group at 
http://groups.google.com/group/g3-5-list?hl=en
Low End Mac RSS feed at feed://lowendmac.com/feed.xml
-~----------~----~----~----~------~----~------~--~---

Reply via email to