seagate st9120817as, very slow cloning

[laptokowiec]

I've got similar problem to
seagate-7200-st3250624a-t10711.html
SMART shows a few hundred bad sectors.
The problem is it scans well in MHDD (shows some UNC and S). Average speed is 40MB/s.
While making the clone (DMDE), copying speed is incredibly slow - 1MB/min.
The same result when using ATOF in MHDD.
Is there a way to speed it up ?

[einstein9]

probably bad sectors or weak dying heads

[laptokowiec]

Then why scaning in MHDD is fast ?

[einstein9]

u will need proper tools my friend to handle this case
+ u will need to know what ur doing

[BlackST]

Maybe it is "like that" with what you have, so "forcefully" you have to stick with it. Behaviour may be different with HW cloning tool OR ...............

Anyway if you can clone it "slowly" but safely, it's better than nothing .

[laptokowiec]

I thought it could be bypassed somehow.

[labtech]

The problem is Windows in combinations with the drive's firmware bug and bads.

In order to maximize recovery efforts, meaning getting a good image out of the drive and of course as fast as possible, you have to work outside of Windows.

Now there are various degrees of effectiveness for each tool as not all can resolve the problem.

What non-windows tools do you have available? Or willing to buy?

[laptokowiec]

I work in DOS scaning(MHDD)/cloning(DMDE/MHDD), so Windows behaviour is not a problem

[einstein9]

I work in DOS scaning(MHDD)/cloning(DMDE/MHDD), so Windows behaviour is not a problem

the way windows detects/reads hdd`s is totally Diff. than Pro. tools

for some hdd`s windows cannot even identify it, but Pro. tools will do

hope u got my point here

good luck

[laptokowiec]

But I've already written that I don't work in Windows.

[Vulcan]

@laptokowiec:

As I understand your description, only MHDD Scan is "fast"; other actions which you have tried (MHDD ATOF, DMDE cloning) are "slow".

There are other tools you could use (Media Tools Pro, or ddrescue under Linux/Unix), which would provide data about which other tests report different speeds, but I see one likely difference from the information you have given so far..: MHDD Scan uses the ATA Verify command; however your attempts to read the disk's data into the host for cloning (e.g. MHDD ATOF and DMDE Cloning) do not use the ATA Verify command (because that does not transfer data to the host!), and instead they must use the ATA Read command.

Therefore if I was in your situation, I would focus on a possible slow host interface, as the cause of the speed difference which you are reporting, since the disk/media is the same in your tests. The Read speed seems even slower than PIO mode (!), but at least I would check to see what host I/O mode had been negotiated, and see if a different mode can be set (e.g. hdparm in Linux).

It's just a thought, based on the information you have given

[labtech]

I am not familiar with DMDE... what kind of tool is this? How does it work?

[fzabkar]

First hit with Google ...
http://www.google.com/search?q=dmde

[laptokowiec]

Vulcan: Thanks. I'll try hdparm.
labtech: DMDE - in my opinion very good software

[laptokowiec]

And what about "Add To Pending" in terminal ?
I think it could slow the cloning.
Is there a way to stop/repair it ?

[drc]

This has all been completely addressed already in the other thread

CE Log error

FW Issue, needs resolving with PC3000 or Salvation.

Ace's UDMA can handle this in the utility settings in DE.

Best regards,

Dobre

It has nothing to do with Windows, or PIO mode, or any of that

[Vulcan]

I wish I had read the linked thread before now - this thread does seem to be a bit of a re-hash

Of course the drive is sick. Nevertheless, I don't see an explanation of why ATA Verify & Read commands to the same media (hopefully to the same LBAs, if the test is being done correctly) appear to behave so differently. While I could conjecture about that (without mentioning PIO etc. ), it would be interesting if anyone actually knows why there is such a difference, as reported by the OP.

[laptokowiec]

I've read it could be solved by p3k or salvation tool, but I still belive' it could be done in terminal mode.
Maybe some sort of 'stop/reset: add to pendind' could be executed?

[laptokowiec]

After 3 hours of TOF in MHDD it suddenly sped up.
I checked the 'pending sectors' in SMART. It reached about 4000.
When I've turned off the system and started it again, the slowness returned.
SMART showed about 40 (pending sectors).
So I figured I'll try the same and see what'll happened.
I run TOF and after about 3 hours it sped up
SMART - about 4000 pending sectors.
It looks that filling the 'pending sectors' position in SMART unsolves the problem.