I've recently had something peculiar occur on one of the hard drives I work with, and I really don't know whether it's okay or bad. I have one hard drive that, when CHKDSK is run on it for surface scan, stays at 0% complete (on the file data scan step) for a very long time, then immediately jumps to 4%. The time it's at 0% is longer than the time it takes to make 4% progress at any point beyond that. This has been observed in the same place on the same drive in two computers.
SMART reports a raw reallocated sector count of 96 (is this the number of sectors, or some internal value that's meaningless to me?), threshold 5, value 100; raw reallocated event count of 135, threshold 0 (none); and CHKDSK and the manufacturer diagnostic tool report no problems.
Is this an indication that that part of the disk is failing, and must be reread repeatedly, or could a very large single file or some other thing cause this (i.e. CHKDSK only updating the progress after scanning each full file)? This is a bit urgent because the warranty expires tomorrow.
You are seeing the beginning of the HD surface going south. If this is a new HD, run a surface scan utility to check the media and block the bad sectors. If the HD > 3 yrs time to backup your data and start looking for a new or replacement drive.
The 4% is a area on the HD that most all of the cat tables are located. I see 99% of HD fail in the 4-5% area.
If you have a copy of SpinRite, use it in maintance mode. It should be able to repair it. I have 1 seagate drive that used 50% of it's reserve sectors right out of the box, and has a bad seek times. I use SpinRite on all the HD I have in use.
Next weird thing: I've got another hard drive that just "crashed". That is, while running CHKDSK surface scan on it, a large number of bad sectors were detected, in several dozen files. However, after CHKDSK finished, and CHKDSK was run again (this time in read-only, structure checking only mode), it said 0 KB in bad sectors. Furthermore, checking SMART data indicated no problems (no relocated sector counts, unrecoverable errors, etc.). What's going on here?