Hard Drives Failing, Who to Believe?

What disk analysis software should i believe?


It first started with windows 10 after creators update going all bonkers. Cant even read USB flashdrives. Checked EventViewer, showed more than 10 errors for, Disk 01 bad sectors, thats most likely my storage drive while my main C drive is an SSD.

So i try out these disk softwares and they all give me varying results. HD Sentinel showing storage drive at 5% poor health. CDI showing unknown, cant read, its site says some drives arnt made for SMART testing or something. SeaTools, has it PASSED.

Then Sandisk, for SSD has it poor. but HD Sentinel that has 5% for D drive, has SSD in perfect health. WTF??? Which is messing up windows. Also why did my flash drive stop working at the same time all this is happening.

I would start with removing the HDD.

The SSD looks normal outside of the odd flag in the Sandisk software.

Most SSDs will show wear-leveling / retired-blocks eventually, and have quite a bit of provision for that.

I couldn't tell you what SeaTools is smoking, but keep in mind that software is basically used as a gate for RMA.

I've been using HDSentinel for a few years now at work, the experience has been positive. It has better detection than CrystalDisk, and puts SMART values into lay terms.

HDSentinel is not a guarantee however, because SMART data is not a guarantee. You're asking a unreliable source (your failing Disk) to perform reliable self-analysis.

i mean HDSentinel says estimated remaining life is 5 days, i guess we will know soon if its right.

It could be a false flag. I once had a drive that would trip smart errors in Crystal and a couple of other apps, but it was perfectly fine years later. However, if the number in HDSentinel is increasing, then yes it's a problem.

it's a seagate.... it could die in 5 secs, no warning what-so-ever!

Something messed up here - I wouldn't trust any program within your current W10 installation, as you said yourself "It first started with windows 10 after creators update going all bonkers"

I would make backups of both disks and install a seperate OS (Linux or Windows 7/8) on a USB stick or seperate SSD/HDD and test disks in question seperatly from there.

Ryokurin said:

It could be a false flag. I once had a drive that would trip smart errors in Crystal and a couple of other apps, but it was perfectly fine years later. However, if the number in HDSentinel is increasing, then yes it's a problem.

Click to expand...

yes it just increased to 4% from 5%.

Folterknecht said:

Something messed up here - I wouldn't trust any program within your current W10 installation, as you said yourself "It first started with windows 10 after creators update going all bonkers"

I would make backups of both disks and install a seperate OS (Linux or Windows 7/8) on a USB stick or seperate SSD/HDD and test disks in question seperatly from there.

Click to expand...

I have Tails, can Tails do that?

Sorry don't know what "Tails" is

Folterknecht said:

Sorry don't know what "Tails" is

Click to expand...

The Amnesic Incognito Live System is a Debian based Linux distro. Hope this helps.

MRFSYS said:

In the situation you have described, we do this:
(a) backup the entire drive, and watch for READ errors
(b) reformat using the long option, examine Event Viewer when finished
(c) scan for and attempt recovery of bad sectors, using OS

The type and number of errors that you will tolerate
are really a management policy question
that you must decide for yourself.

Click to expand...

This is good advice. I had a freenas setup for a while that kept saying I had harddrives going bad. I swapped out cables, moved drives around, re-created pools and it kept saying via SMART there was issues with several drives. I almost tossed them into the trash but then decided to do a complete sector format and stuck them in another box and they pass SMART with flying colors.

SMART issues should never be ignored but they don't always mean toss the drive. Case by case it.

So the manufacturer tools are always the last to check. They are going to be the least conservative since they are what is used to determine if a drive is RMA'd or not (in the case of Seagate). So to them, a disk full of bad sectors is still 'healthy' as long as it can continue reallocating.

You have to make a judgement call and look at what the errors being reported actually are and determine then if you should replace the drive.

If in doubt, file a RMA request with the manf. Send the evidence from all the tests. Even if they deny at the time, you have a documented record of the problems while the drive is in the warranty period. Years ago, I had a drive that was part of an array fail. Swapped in the spare which worked fine. Initial tests on the failed drive with the Seagate tools showed fail but website said follow procedure X(don't remember all the steps) before filing RMA. Procedure X cleared the errors. Put drive on shelf. Couple of years later, tried to use the drive. OS installed but kept failing with a lot of drive noise. Turns out drive WAS bad. Of course, warranty period was over with no official record of problems during warranty period.

Canvas said:

i mean HDSentinel says estimated remaining life is 5 days, i guess we will know soon if its right.

Click to expand...

HDSentinel is really designed to be run either all the time (which I think is overkill and ridiculous) or at least once per week or month (I try to run it once in a while manually) to check on things and keep tabs on how things are changing over time.

Running it once, especially after the drive has been experiencing issues for a while, and expecting it to give you things like accurate days left and percentages of life remaining are unreasonable, IMO.
HDSentinel has no prior history of your drive, so it does not know how long it took for it to get to the state it is in now. Did all of those SMART errors happen in one evening last week or were they spread out evenly over the hundreds or thousands of hours that the drive has been running? SMART data does not keep a log history, only the current state of each parameter. Each time you run HDSentinel it compares how the drive looks now with how it looked from the previous times you ran the program. This gives it trend analysis over time (you'll see line graphs appear in the program window) which can be a helpful indicator of how the drive is aging.
If there's no history for the drive, HDSentinel can't do anything but throw up alarms saying "hey, this thing is bad and could die at any moment." Conversely, if it said you had 3 months of life left and then the drive died 10 days later you'd be pretty pissed, I'm sure, so they err on the side of caution when there is no SMART history recorded.