How much of an ancient method this is compared to a program like Power Quest
Drive Image. It creates an image of a partition on the second drive and the
ONLY requirement is that there be enough space fro it to fit on the drive in
the target partition. It will even build the correct partition type if the
target is big enough.
No problem , No loss.
Of course I don't know if there is a version for Linux, would be nice!
Oh yes, 115 Mb/Per minute backup and 220 Megabytes per minute restore !
>-----Original Message-----
>From: Charles Curley [mailto:[EMAIL PROTECTED]]
>Sent: Saturday, July 29, 2000 6:10 PM
>To: [EMAIL PROTECTED]
>Subject: Re: [expert] copying whole harddrive !
>
>
>On Sat, Jul 29, 2000 at 02:55:44PM -0700, Dave Lers wrote:
>> On Sat, 29 Jul 2000, Charles Curley wrote:
>> > On Sat, Jul 29, 2000 at 11:04:03AM -0700, Dave Lers wrote:
>> > > A simple
>> > >
>> > > dd if=/dev/hda of=/dev/hdc
>> > >
>> > > works great, super backup method, all you need to do is swap drives
>> > > (plug hdc in where hda was). You want both drives as master, 1 per
>> > > channel. The drive you are writing to needs to be the same size or
>> > > bigger. Since my drives are almost identical I haven't played with,
>> > > don't know how to, utilize extra space if copying to a noticeably
>> > > larger drive.
>> >
>> > This assumes that the two drives have identical sectors per track and
>> > tracks per cylinder. Otherwise the drive geometry won't match
>the values
>> > in the MBR's partition table. If the new drive has larger
>values, you are
>> > OK, but you waste a lot of space. If the new drive has samller
>values, you
>> > are hosed.
>>
>> I don't know anything about drive geometry, are you saying there is
>> more to it than making sure the copied to HD is >= in size? How much
>> wasted space are we talking about? Are you talking about the
>> leftover at end of the disk?, which I assumed could be formated as
>> another partition.
>
>Well, you've just opened up a nice big can of lawyers, so sit back and
>read.
>
>There are three elements of a hard drive's geometry: number of sectors per
>track, number of tracks per cylinder, and total number of cylinders. I'll
>ignore the sector size as almost all hard drives leave the factory
>formatted with 512 byte sectors.
>
>OK, suppose the old hard drive has 35 sectors per track. If the new hard
>drive has fewer, say, 30, then dd will write the 31st sector of a give
>track somewhere else, Murphy only knows where, or it will error out and
>refuse to write it. So you are hosed.
>
>If the new drive has more sectors per track than the old one, one of two
>things will happen when dd writes to the new one. Either dd will be
>"smart" (i.e. copies using the old drive's geometry) and ignore the spare
>sectors, in which case you have a working hard drive, and waste those
>spare sectors. The spare sectors are spread out on each track all over the
>new drive, and so they aren't recoverable in any useful way.
>
>Alternatively, suppose dd is "stupid" (writes the data using the new
>drive's geometry). Now, the first track will be copied fine. But the first
>N sectors of the second track (where N is the difference in the sectors
>per track) will be copied from the second track on the old drive to the
>first trackon the new drive. But, given the assumptions built into thte
>partition table of the MBR, that ain't where the file system driver is
>going to look for them. So you will get a lot of information showing up in
>the wrong place.
>
>I have no idea whether dd is "smart" or "stupid". If I were writing it,
>I'd make it "stupid" because most of the time one copies other things
>besides hard drive to hard drive, so the smarts would be irrelevant.
>
>Now, think the same thing through for the number of tracks per
>cylinder. Again for the total number of cylinders.
>
>Also, even if you end up with spare cylinders, you may not be able to use
>them. You have to have at least one primary partition available that you
>can use to build an extended partition with the new cylinders.
>
>Also, as far as I know dd does not manipulate either the drive partition
>table in the MBR or any information in any of the file systems. It is
>strictly a raw copy, byte for byte. So your MBR on the new drive will most
>likely be incorrect.
>
>Have I gotten you totally confused yet?
>
>--
>
> -- C^2
>
>No windows were crashed in the making of this email.
>
>Looking for fine software and/or web pages?
>http://w3.trib.com/~ccurley
