Re: SCO 5.0.5 dd / clone
- From: "Otto" <otto@xxxxxxxxxxxxxx>
- Date: 8 Jan 2006 14:50:04 -0800
Hi Bela,
Thanks for the reply. Indeed I tried it with both the old wd driver (as
originally delivered in 5.0.5) as well as the newer wd driver as found
in 5.0.7. The timing is DRASTICALLY diff (the old wd driver took almost
4 hrs to complete the clone whilst the new wd driver finished in under
25 minutes for a 20Gb clone!!). However, the result seems to be the
same - the new clone is badly corrupted. As I mentioned previously this
only happens on a "live" running system. If I boot from say a CD and
clone drive 1 to drive 2 it works just fine. Again, any help or ideas
to try would be much appreciated. Rgds. Otto.
Bela Lubkin wrote:
> Otto Rodusek wrote:
>
> > Without going into a lot of detail and reasons for backup decisions, I
> > have implemented a dd cloning of the master hard disk to a backup disk.
> > That is to say my customer has 2 hard disks (in a caddy), the primary
> > or master disk (hd00) and the clone or backup disk (hd10). A cron job
> > wakes up twice daily and does a complete clone of the master disk to
> > the backup / clone disk using dd. No great issue here as the entire
> > disk (20Gb) takes about 24 minutes to complete. Now in Linux, this
> > works great. If the master disk crashes for any reason, the customer
> > simply re-boots their system and boots from the secondary drive whilst
> > the primary is being repaired and all boots up fine and proper and no
> > problems.
> >
> > However, the same process in SCO 5.0.5 seems to generate a completely
> > corrupted disk. When I compare the master to the clone, you can see
> > that the clone is completely a mess. I don't understand why this is the
> > case with SCO but not with linux. As I'm doing a dd of the ENTIRE disk
> > (a binary bit by bit copy), I don't understand why it doesn't work
> > under sco. BTW - in SCO I don't boot from the secondary drive - I have
> > to set it to the master drive first however, as per above the disk is
> > so badly corrupted that the boot process will not work.
> >
> > The command I use is
> >
> > time dd if=/dev/hd00 of=/dev/hd10 bs=1451520 count=12529 (for SCO)
> > time dd if=/dev/hda of=/dev/hdb bs=1451520 count=12529 (for linux)
> >
> > Some additional info. If I use a "live cd" to boot and do the above
> > clone for sco it all works fine - hence I confirm that my drives and
> > controller are healthy. The problem only occurs only if I boot the HDD
> > (primary disk) and then try the dd clone. So obviously dd doesn't seem
> > to like doing something on a running system but not sure why this
> > should be so. Again, under a running Linux system this works perfectly
> > fine.
>
> Install the newest "wd" driver you can, try this again. There was a
> data corruption problem much like you describe, happening only under
> certain kinds of two-drive simultaneous loads.
>
> Hopefully someone else will chime in with the name and location of the
> newest "wd" driver known to work with OSR505. This is not necessarily
> the very newest "wd".
>
> I am not 100% certain that this problem _can_ be fixed under OSR505. I
> don't remember if the fix was in "wd" alone or involved other parts of
> the kernel. You might need to put OSR507 on the machine. (Note that
> when OSR505 first shipped, the idea of a 20GB IDE disk was some sort of
> bizarre fantasy...)
>
> >Bela<
.
- References:
- SCO 5.0.5 dd / clone
- From: Otto
- Re: SCO 5.0.5 dd / clone
- From: Bela Lubkin
- SCO 5.0.5 dd / clone
- Prev by Date: Re: ATI Radeon 7500
- Next by Date: Re: ATI Radeon 7500
- Previous by thread: Re: SCO 5.0.5 dd / clone
- Next by thread: ATI Radeon 7500
- Index(es):
Relevant Pages
|
