chkdsk 1st corrupts, ultimately fixes after many runs

B

Brister

Problem:
On bootup chkdsk decides the file-structure on a drive
needs fixing. chkdsk runs and corrupts most files
working fine before bootup. Upon reboot chkdsk may or
may not decide file-system needs further fixing, but on
subsequent reboots will run at least once, sometimes as
many as 5 times (so far). The affected drive is never
the boot (C:) drive, but always one of the other ones.

The perfectly fine file system is always corrupted by
this initial chkdsk. At some point after the 2nd, or up
to fifth time chkdsk runs itself the file system will
actually be put back to where it was before the initial
chkdsk operation. The subsequent system initiated runs
of chkdsk typically don't happen with subsequent reboots,
but will often skip a boot and then self-run again the
next time.

Running chkdsk from the GUI drive\properties does not fix
anything and will come back "clean" if run after this
initial chkdsk is run. For some reason the further
chkdsks only do something when the system decides the
file-system is dirty upon boot.

This problem tends to occur when I do a fresh
installation of the Operating System. I periodically do
fresh installations of XP on the same machine because I
have found the system can become corrupt over time and I
have not found any tool that can fix the system
properly. Even most of those "doctors",
registry "repair" solutions, fix/clean/whatever super
utilities might fix one thing, and then go screw
something else up. So I have found the system only runs
well after a CLEAN INSTALLATION, with all SERVICE PACKS
and a VERY CAREFULLY SELECTED set of applications and
utilities chosen from years of experience.

The event log says there is a problem with the drive
hence the chkdsk run. I might be inclined to believe it
if it weren't for:
a) the drive was working just fine before chkdsk and then
was totally screwed up after chkdsk got the bright idea
the drive wasn't working and totally messed it up. who
ya gonna trust?
b) it has occurred on different drives in the same
system, which might point to a controller problem, except
c) has occurred on drives connected to different
controllers, including RAID drives, which might point to
some sort of generic hardware problem convincing chkdsk
that the drives are at fault, except
d) all diagnostic software I've run has come back clean
and when did chkdsk get so smart that it'll know
unrelated hardware issues, which no other software can
find, are corrupting the file systems in the drives, and
e) chkdsk is completely wrong! the file systems were
fine and IT goes and messes them up.

Since the system tends to fix itself, one might ask why
worry? Worrying: a) if any of XP's automatic
default "disk optimazation" or other system
utility "decides" to run then the file system will
actually be totally messed up, and b) if I forget which
drive chkdsk is messing with operator error could botch
thngs, c) its just plain bogus, somethings going on here.

System:
XP Pro, might have all service updates, might have some,
or might have none depending upon when chkdsk decides to
go fix things.
HW: Hyperthreading P4 2.6Ghz; 875p chipset with Intel
ICH5 IDE and HighPoint 375n RAID on DFI motherboard; 2GB
Corsair matched DDR memory (4 modules) running at 200Mhz
effective 800Mhz FSB; ATI Radeon 9700 w/128MB; currently
7 hard-drives: 1 120GB WD boot/system drive (never falls
prey to chkdsk), 1 200GB WD drive (has fallen prey maybe
6 times), 1 250GB Maxtor drive (just finished his 4th
encounter with chkdsk, took 4 system-initiated chkdsk
runs to put things back), 2 200GB WD drives configured as
a RAID 0 400GB drive (first drives to experience chkdsk's
wrath, but only hit once), 2 250GB forgot-the-brand
drives configured as RAID 0 500GB drive (so far have
flown below chkdsk's RADAR, crossing fingers),
periodically other Maxtor, WD, Quantum, etc. drives are
connected via IDE or USB for extended periods wihtout
incident; Pioneer 104 DVD-RW; generic 1.44 floppy; all
running in a Thermaltake SuperX case with 14 fans and a
450w PS running (entire system consuming 220w according
to UPS); generic mouse, generic keyboard, generic monitor.

I suppose I could start replacing random pieces of
hardware until the problem goes away. The problem
doesn't occur with enough frequency for this solution to
be readily ascertained for effectiveness, yet occurs with
enough frequency to really p*ss me off, so I'm not sure
this is viable and I think its always wise to know what
the actual PROBLEM is before you start operating.

So if anybody knows of a solution I haven't figured out,
an MS KB article I haven't read, or has any other ideas,
clues, or visions as to what might be happening and what
I might do to fix it I would be most appreciative.

Or, is XP Pro just not up to the task of managing my
hardware set? If that's the case then we're all in
trouble because banks are using this stuff to process our
checks. Yikes!

Misc:
Usually occurs, though not always, on a fresh OS
installation.
Not running zone alarm.
Not running any drive keeper or other such software.
Problem can occur on a clean XP install.
No hauntings or other paranomral activity "sensed" in the
area.

THANKS IN ADVANCE!!!

Direct reply to e-mail after removing the stuff up to
the "b" and after the "0" for the e-mail "name".
 
B

Brister

Mary,

I don't want to seem ingrateful, and I do appreciate that
Microsoft folks cruise Usenet, however your response was
not only profoundly inappropriate, but showed a genuine
lack of concern that is so rare these days.

Be that as it may, I followed your advice and am no
further along than before. The only potential piece of
information I gleaned was a mention of some log-file
being corrupt. Further searches for said log-file and
potential File Table corruption proved useless.
 
R

RJK

Having a slightly more elaborate system than most, perhaps you need to apply
a slightly more elaborate and methodical approach. A thought that crossed
mind, during the umpteenth read of your post, (because it was so big
.....just kidding:), was, "could there be a dodgy IDE ribbon, or ribbons, in
there. Assuming that NO hardware, anywhere is faulty :-
Another thought that crossed my mind, was, "have you downloaded the hd
maunfacturers hd diagnostics for each drive and run, on each drive, to rule
out a faulty drive. "...forgot the brand," indicates that you're not being
methodical enough and have not run hd manufacturers diagnostics on each
drive.
And how about memory, ...looks like you're running the fsb flat out - how
about reducing it to 166mhz just as a test to see if the problem goes away.
i.e. could the problem be along the lines of faulty memory where the
registry gets corrupted. I appreciate that you're not getting a "problem
with the registry - Windows'll fix it and reboot."

What about IDE speeds e.g. 133mhz data transfer on drives other than SATA
150mhz, ...could there be something going on there ? How about setting all
non SATA drives to UDMA5 or less. Have you got 5,400 and 7,200 rpm hd's on
the same ribbon somwhere? i.e. is there a hard disk / brand incompatibility
in there somewhere. I ask because a while ago, I had two drives, (different
brands - WD and either Fujitsu or Seagate ...if memory serves), on the same
80 core IDE ribbon and bios hd auto detect set them at UDMA 6, (I think it
was), and I had a file handling problem in Windows - deleting a file or two
would take up to a couple of minutes !
When I set both hd geometries to 100mhz / UDMA 5, the problem went away. I
quickly replaced the drive that claimed to support ATA133.
....anyway just a couple of thoughts :)

regards, Richard
 

Ask a Question

Want to reply to this thread or ask your own question?

You'll need to choose a username for the site, which only take a couple of moments. After that, you can post your question and our members will help you out.

Ask a Question

Top