[prev in list] [next in list] [prev in thread] [next in thread] 

List:       linux-raid
Subject:    Re: Partition table disappearing
From:       Rob Hagopian <hagopiar () vuser ! vu ! union ! edu>
Date:       1997-10-03 4:08:40
[Download RAW message or body]

I just tried it on a whole bunch of drives:
1 4G Caviar on a Dual PPro did show what you described
2 4G Caviars on a P120 did NOT show what you described (1 was created with
  the exact same commands as the one on the PPro)
9 SCSI drives (only one partition on each though) did NOT show any
  problems
3 IDE drives smaller than 500MB

I believe that the problem may be something to do with LBA and conflicting
schemes of accessing large drives.
								-Rob H.

On Thu, 2 Oct 1997, Linas Vepstas wrote:

> I too have frequently seen fdisk choose what looked like bizarre
> start/begin/end values. At times, i've seen mutliple cylinders
> overlap (like hundreds) between end of last and begin of
> next (but not start of next).  I always viewed this as rather
> alarming, and once spent half a day toggling between different
> display units, mutiplying big numbers in my head, erasing and
> re-creating partitions.  In fact, my fdisk almost always
> generates error messages about the partitions that it created,
> claiming that there is someething wrong.  Try it .. use the "verify"
> function on fdisk. You'll get errors, no matter what, and its impossible
> to create partitions that don't generate errors.   I've tried using
> newer/different versions of fdisk, all with the same result.
> 
> I find this behaviour very alarming, and since I've installed many
> linux machines now, with all sorts of brands and types of disks,
> I just sort of ignore it.  I usually just plow ahead in frustration,
> and, for what its worth, have never ever seen an actual breakdown
> of any sort as a result of this.  So I close my eyes and move on.
> 
> However, fdisk should be fixed so that it behaves in a rational manner.
> The bad interaction with RAID is the first concrete evidence of actual
> damage caused by this brokenness.
> 
> --linas
> 

[prev in list] [next in list] [prev in thread] [next in thread]