Crono said:
A viable alternative to figure out which is happening is to use a an NTFS driver on a linux system (NTFS-3G) and use one of their many filesystem checkers. It only analyzes. It will tell you if this is the case. (They also generally have far better S.M.A.R.T. utilities which will give you a pretty good idea of the stat of the hardware)
Seatools for DOS gets the job done (smart for Maxtor/Seagate Drives and bad sectors for all) - theres also PC doctor, but only for the hard drive test. I constantly use Hirens Boot CD for all my tools. There are a ton of tools on there, it even comes with a live windows XP OS so you can boot and tinker with the files (delete virus files, repair the registry, etc). There are some southbridges that it doesn't work with, but it works 90% of the time.
If you need to check system files (different from check disk) you can use microsoft's tool for that ... I forget the name of it, syschk? That'll check against system files integrity.
"sfc /scannow" is what your referring to methinks
But that only verifies all critical windows files are present and pass a crc checksum (or whatever windows uses).
The majority of time, you have to repair the actions checkdisk takes because it renames things and moves them around ... even if there is no corruption. The majority of time it's ran because it's bundled with various other actions (windows defragmentation) or from a incorrect shutdown.
I just don't see the point of using software that makes more messes then it fixes when the messes it's made to fix aren't usually fixable anyway. (Especially considering you could make a live CD that has all the needed utilities that you boot into when you need to check)
Problem is your average user doesn't know how to run live CDs, let alone know how to type:
mount -t ntfs-3g -o force /dev/sdb1 /media/Drive
into the terminal (Ubuntu style, not your normal linux style, I just use ubuntu because it is the most common distro). Even the automated live CD drive mounter scripts have problems.
I only ask because I've used chkdsk multiple times in my computer repair line of work:
-Computer gives the "blue screen" (not an actual blue screen) of "cant find windows\system32\config\system". Chkdsk finds the registry hive and moves it and fixes the problem. We then move onto cloning the drive (acronis or ghost) to a new one.
-(different case than above) Cloned drive wont boot because the original drive was failing. Ran a full chkdsk. Boom fixes the problem.
-(data recovery), data wont copy (drag and drop) nor will it be recovered by commercial software. Run chkdsk on drive (over-night), come back can drag-and-drop data off drive OR commercial data recovery software gets to the data (multiple separate occasions).
Chkdsk does a lot of low level sector manipulation for sure, thats why if you stop it before it finishes its going to cause problems. Any non-readonly repair tool is going to do that.
I guess the debate is whether chkdsk should have come or not with windows. I'm glad it did, but its meant to be used as a tool, not a cure-all like the way it runs.
Orph- That drive was failing already. And windows was detecting that, it was attempting to fix the problem automatically with a chkdsk scan. It wasn't necessarily chkdsks fault that it killed your drive. Failing drives are fickle, and a heavy operations such as chkdsk pushes it over the edge. Sounds like im proving my own point with circular reasoning, kinda, but even if you ran another "magical" ("magical" because I don't know of any that really exist that repair ntfs file systems) tool that fixes poor sector problems it would have damaged the drive just as much as chkdsk did.
too-long-didn't-read: chkdsk should be used as a tool by someone who knows what they're doing, in its current state its relied on my M$ and users as a cure-all, when it shouldn't.