home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.sys.mac.system
- Path: sparky!uunet!xstor!NewsWatcher!user
- From: daniels@xstor.com (Daniel A. Segel)
- Subject: Re: HUGE disks and minimum file size
- Followup-To: comp.sys.mac.system
- Message-ID: <daniels-121192104732@175.175.10.2>
- References: <bruce-b-061192140954@130.216.34.254> <1992Nov6.194833.6234@bnr.ca> <RMF.92Nov9143202@chopin.cs.columbia.edu>
- Organization: Storage Dimensions, Inc.
- Date: Thu, 12 Nov 92 19:15:25 GMT
- Lines: 70
-
- In article <RMF.92Nov9143202@chopin.cs.columbia.edu>,
- rmf@chopin.cs.columbia.edu (Robert M. Fuhrer) wrote:
- > I also think that the user should have some control over the min allocation
- > clump size when partitioning. It's crazy to *have* to break a nice big drive
- > into a whole bunch of small ones just to avoid wasting *large* amounts of
- > space. As a better, long-range solution, a more reasonable allocation policy is
- > in order. [Apple, are you listening?]
- The operating system needs to store the number of logical blocks
- (allocation blocks) in some form of variable. Apple chose a variable that
- was only 16 bits wide, so you can only have up to 65,535 allocation blocks
- on a hard drive. If the number of physical blocks (i.e. 512-byte or
- 1024-byte sectors) exceeds this amount, which will happen on any drive over
- 32MB, you need to allocate logical blocks equal to some multiple of
- physical blocks (i.e. for a 64MB drive, make each logical block 2 physical
- blocks, or 1024 bytes).
-
- It would require a fairly big re-write of the operating system to suddenly
- change over to a different system of keeping track of physical blocks on a
- SCSI hard drive, but it could happen.....
-
- >
- > Now, for a follow-up:
- >
- > After the re-partitioning, I soon had a machine crash which damaged the boot
- > partition. The system booted to a floppy-question-mark icon (no bootable
- > volumne), and after booting from a floppy, said the drive was unreadable.
- >
- > I have never before in 5 years of running this Mac II, 1 and 1/2 years of it
- > with this drive, experienced a crash which took the hard drive with it.
- I work in tech support for a hard drive company. We get upwards of 100
- calls a week from people whose hard drives have crashed, and about 10-20%
- of those calls involve a corrupted partition map. If you have software that
- knows how to recreate the partition map, this isn't much of a problem.
-
- Unfortunately, most software can't. In fact, most people don't know much
- about partition maps, and when the partition map gets corrupted utilities
- like Norton and MacTools tend to just say "there is a problem I can't fix"
- and give up, when in fact there may be nothing wrong with the directories -
- it's just that everybody is looking for them in the wrong place since the
- partition map is pointing to the wrong place.
-
- > Now, my question is this: is a partitioned drive on a Mac more susceptible to
- > being damaged than a non-partitioned one? [I can think of several reasons why
- > this might be the case, but I'm looking for empirical answers, not theoretical
- > ones.]
- There is a partition map on every Mac SCSI hard drive, and whether it has
- one partition entry or twenty is irrelevant. The partition map can be
- corrupted just like the directories, since it is just a bunch of data
- sitting on the hard drive, albeit it is always in a specific location
- (starting at block #1, going for upwards of 64 blocks, depending on the
- formatting software used).
-
- I guess one could say that the more entries there are in the partition map
- the more likely that random garbage written to that part of the disk will
- damage something, but I wouldn't say that makes single partition drives
- safer...
-
- One last thing: recovering data off of a crashed drive is always easier if
- the utility used knows where to look for the directories; with more than
- one partition there is a greater chance that the directories of the
- additional partitions will not be locatable...
-
- Daniel S.
-
- ___________________________________________________________________________
- Daniel A. Segel, KD6NMT Macintosh Customer Engineering
- Triumph Bonneville 650 Storage Dimensions, Inc. (408) 954-0710
- daniels@xstor.com All opinions are of my own invention.
-
-
-