"Disk error..."

Chas:

= I randomly get a blue screen with the message "Disk error. Cannot write to
dis
= in drive D. Data might be lost or corrupted. Press any key to continue."
I
= ontinue with no corruption or data loss. There is no eror code. 2 HDD C i
ma
= ter, D is slave. Installed all W98SE startup and shutdown patches. Reseate
me
= mory, new 600watt power supply(happened before I changed the P.S. Dr.Watson
rep
= rts no problems during occurence. Used Seagate tools to diagnose under com
and
= prompt,everything is fine. No conflicts in device manager. Looks like a har
war
= error but how to pinpoint?

Check for a slightly loose data connector, or possibly power
connector. I had something similar years ago except only one hard
drive. Sometimes boot, sometimes not. Almost had to use a pair of
pliers to remove the data cable at the hard drive side but could blink
to move the able at the motherboard/controller side. Replaced the
cable, no more problems.

-
¯ barry.martinþATþthesafebbs.zeppole.com ®

* BeerReader Ver. 16 oz. þ When In Doubt, Assume It's Funny.
---
þ RoseReader 2.52á P003186
þ The Safe BBS þ Bettendorf, IA 563-359-1971
---
þ RIMEGate(tm)/RGXMod V1.13 at BBSWORLD *

Hi,

As you have anticipated. Yes, it is a hardware problem. i believe
there are few bad sectors on you disk. to check it run Scandisk from the
command prompt or from run.

after opening scandisk seclect thourgh scan and run complete scan. it
might take lot of time depending on the size of the disk. scandisk will
point at any error in your disk and fix it. you will have to perform
scandisk on all the partitions.

I am happy to help you, let me know about your progress on this issue.


Harish.G

This posting is provided "AS IS" with no warranties, and confers no rights.

On Wed, 16 Jun 2004 08:35:15 GMT,
(Harish.G [MSFT]) wrote:

Another one, from an earlier thread?

If so, this dude already has the click of death. There's no wishing
that away with Scandisk, but there's every risk that protracted and
pointless Scandisk sessions to kill the HD stone dead.

As you have anticipated. Yes, it is a hardware problem. i believe
there are few bad sectors on you disk. to check it run Scandisk from the
command prompt or from run.

There's no quick overview mode that will tell you wheher you have
"just one bad sector" or massive widespread surface deterioration,
possibly worsened by an abrasive airspace (from desbris after a head
strike) or failing head positioning mechanism.

Scandisk won't have a clue. It "fixes" errors one at a time, as it
finds them. If every sectors's bad, it will try and "fix" every
sector - it has no overview awareness to stop and say "Hell, this is a
mess! What am I doing here? Abort NOW and evacuate your data!!"

Scandisk can't read a failing bad sector any better than a Copy
operation can. What it can do is retry it again and again, hoping to
get a clear shot at the data (as "proven" by a CRC that passes).

But this is exactly what modern HDs do behind the scenes, anyway.
When a Copy operation tries to access disk that won't read easily, the
HD's firmware springs into action to "fix" the bad sector. It
attempts to read the contents of the sector, writes this to a "spare"
sector, and fixes up the internal raw addressing to point to the
working "spare" sector instead.

All of this happens beneath the file system's level of abstraction, so
you don't see any B(ad) blocks in surface scan, etc. This has prolly
happened many times already; suspect this when you see an
ideosyncratic slowness with HD LED on and mouse pointer stops moving.

Scandisk's surface "fixing" is the same idea, but cruder, and
operating at a higher (file system) level of abstraction. Where data
is concerned, the file system works in clusters, which contain a
number of sectors. Scandisk "fixes" a bad cluster by copying readable
sectors from that cluster to another cluster, then updating the FAT or
directory entry to point to the new cluster instead. The old (bad)
cluster is then marked as Bad in the FAT, so it won't be used, and
will show up as a B(ad) block on surface scan (which won't re-test it)

In all of the above, there's NO promise that the data read from the
bad sector will be recovered properly, with errors, or replaced with
pure garbage (say, all zeros, all FFh or those divide-by symbols so
beloved by format routines). But once the problem has been "fixed",
it can't be re-detected as a problem. You just threw the needle back
into the haystack - let's hope you Appended the log?

after opening scandisk seclect thourgh scan and run complete scan. it
might take lot of time depending on the size of the disk. scandisk will
point at any error in your disk and fix it. you will have to perform
scandisk on all the partitions.

Do NOT do that before evacuating your data. Priorities! Raw money
will buy this dude another HD, but there is NOTHING you or I can do to
get his data back if the dying HD eats it.

Sorry to sound harsh, but I've seen ppl in tears after generic IT
techs have utterly and irreversably botched data recovery situations.
"Scandisk said it couldn't fix it, so we formatted and re-installed
Windows" - that sort of criminally negligent garbage.



-------------------- ----- ---- --- -- - - - -
No, perfection is not an entrance requirement.
We'll settle for integrity and humility
-------------------- ----- ---- --- -- - - - -