External spinner drive CHKDSK strangeness

I ran chkdsk /f and it looked ok.
I deleted about 120GB of ATI images.
Then overnight I ran chkdsk /r because previously, in the event logs, I saw disk controller error (event 11) and bad block (event 7). And as I started chkdsk /f event 7 occured as well.
Result surprised me: Windows allocated over 50% of the drive to bad sectors and reports zero bytes free.
What remains on the drive seems readable and Ok.

Did big deleting cause it?
Should I give more info? if yes, please ask.
I just want to learn. Drive is probably toast.

 

While not exactly to my element in offering the advice you might be looking for, others will weigh in for you on that. And thanks for the supporting screenshots.

AFAIK the Windows chkdsk program has proven reasonably accurate enough in it's findings. I've battled with bad blocks and such before on spinners and even after a clean sweep wipe then reinstall, still ended up with HDD that was all but toast. It happens that way sometimes.

But wait on a more informed expert who can better lay it out for you from solid knowledge of these matters.

Thanks

 

Thanks, EASTER.
Here's a funny part: I ran HDsentinel's read only surface scan for 2 hours, and 10000 blocks are all perfect. Go figure.
At least from HDsentinel I learned what a "block" is
In the chkdsk statistics they added some 75 million bad clusters from free area(s). Hmm, what's bad clusters me wonders.

 

Just curious, why did you run checkdisk in the first place? Was there an actual issue that prompted you to do so, or were you just performing a maintenance task?

 

@act8192 Did chkdsk reported bad clusters or did it only recover orphaned files? (Check for a folder named "Found.xxx", eg. "Found.001").
Also check the cable and port of the esternal drive (low power and bad cables can cause disconnections and can corrupt the files table if you transfer files at the moment it happens)

Panagiotis

 

Hey Liz,

I do not think that your drive is toast. This kind of stuff can happen if there is a controller error, or as pandlouk states if there is a bad cable or a power supply problem or whatever hardware problem.

The remarkable thing in your second chkdsk report is that the complete user area of the drive is OK, but each and every free cluster is marked as bad. This is highly unlikely, and if HDsentinel can read every block without errors then you got a very good chance to recover the drive completely.

Of course before you start experimenting you should backup all the files from the drive to a different location. Then I would try to wipe the free area with some HDD wiper software (CCleaner has such a tool under Extras). And then you need to reset the Bad Sector count. For an NTFS partition you can use "chkdsk /B" for this.

If this does not work then you might have to use the HDD diagnostic software from Toshiba (or use the WD Data Lifeguard Diagnostics tool) to perform a so-called Low Level Format of the drive. This will write zeros to every sector of the drive giving the firmware a chance to remap bad or weak sectors. Now partition and format the drive, and the bad sectors should be history.

Good luck
manolito

 

Why chkdsk? I had trouble with safe removal and had to restart windows to close things up, and had trouble validating two ATI images with ATI running in boot mode not in windows. That's the 2 images I referred to in post 1. So I ditched them.
I snooped the event log, found events 11 and 7(see post 1), and later on, NTFS event 137 which said "default transaction resource manager on volume N: encountered a non-retryable error and could not start" whatever that means.

 

- No idea if it recovered orphaned files. Definitely reported bad clusters when it messed with free space (that's how I read the second chkdsk shot in post 1).
- If I recall correctly from XP times, found.xxx files would be on that external disk, right? No such there.
- I don't know how to check the cable. Removing it from the drive is not possible or maybe too tough for me.
- Low power - unlikely. Even if street power goofs up the battery would take over and , I think, supply sufficient power to a small drive connected to the usb3 port directly, no hub. What do you think?

 

Cool. Great information and tips about new tools. In my opinion the drive is shot (controller error?) and I can do what you said to see how it all works. It'll be an excellent experiment and then watch the drive over time without trusting it
What is a controller error - it's inside that drive, right?
But why would Windows declare so much as bad in the free space?

I always wondered what that wiping of free space was about. Thanks.
I don't really care about the remaining ATI image, and the Syncback data I have on another 2 drives anyway.
Will that chkdsk /B run as long as chkdsk /r did, i.e. 4 hours?

OK. Windows allows to do quick format and slow format. Is slow format the same as the Low Level job you mention?
Curious thing is now that SMART in HDSentinel reports no reallocated and no bad sectors.

 

Maybe running the HDsentinel check already did reset the bad sector count. I found this in the HDsentinel forum:

If you are hard core maybe have a look at this 10 year old thread at Doom9:
http://forum.doom9.org/showthread.php?p=1339304#post1339304

DFSee is still one of my favorite tools for HDD problems.

 

HDSentinel may have reset the count because SMART shows no errors. I doubt it though. I wish I saw SMART values before all the trouble started. Even if they did, Windows still refuses to release those bad sectors.
I will likely do that Reinitialise Disk surface or chkdsk /b. I went to the HDsentinel forum and saw similar conclusions to your quote.

I'm not hard core anything, just muddling through. That doom9 info is a very interesting thread. Thank you

EDIT: To Reinitialise in HDSentinel requires Pro version which I don't have.
Running CCleaner as Admin to wipe free space was willing to work, started with MFT and aborted with not enough space. Arrgghh. So back to figuring it out.

 

@act8192
you cannot change the usb cable? (it's attached on the drive?)

format your entire drive and uncheck the option "perform a quick format" (It will zero out the surface) or better with hdsentinel run a write test on the entire surface.

a controller error can ba caused also by a bad driver.

 

I was just referring to the DFSee software. You need to be "Hard Core" to use it.

As far as I remember resetting the bad sector count with DFSee was very fast. Using "chkdsk /b" instead may take some time because it includes the /r option.

If you can get your drive back to normal after a "Low Level Format" then using chkdsk without the /r parameter should not show any bad sectors. But if you restore an image based backup which was taken with the bad sectors present then the bad sectors will be there again. This is the situation where you want to reset the bad sectors count.

If you want to try DFSee let me know. It really is intimidating, even for me, but I can show you how to perform the bad sector reset quickly without hassle...

 

The cable is not attached. I remember plugging it in. But to remove it I may have to use pliers or a wrench or something. It just won't com out.

After seeing this and Manolito's old post, I did some googling and found that slow format will write zeros, and it did Thank you.

Controller - driver? Might be. it's MS issue 2006. They put it in. Considering that I have two other Convio drives which don't complain about controller, I'll probably leave the driver as is since driver things scare me.

 

I did slow format. Two hours later the drive was empty, 465GB free as it should be. I looked at it with some hex editor and, indeed, it is filled with zeros. Except for few, 10 or so sectors, probably System volume info.
I did some copying and deleting. So far so good.
I ditched all ATI images, so nothing bad would restore
I don't trust this drive. SMART in defraggler is practically identical to what it was before, but before hasn't showed errors. And, as I complained in post 1, chkdsk didn't either until that second screenshot. So it's all weird in my mind that things don't report bad sectors at all till you run chkdsk /b.

I'll take a pass this time. Let's see how it'll hold up. It might not be worth it.
I plan to run chkdsk /b on it, perhaps tomorrow. I already saw one controller error when I ran that hex viewer, possibly just before as I plugged it in, but no other errors.

 

You are welcome.

From you description (cable stuck) I would not trust it unless you manage to remove it first... either the cable plug or the disk's socket are bend and unless solved I would not use it for important data.

Leave it unplugged from the pc at least for 15/30 minutes.(it could be stuck because of heat expansion of the plug and a tight socket) and try removing it with gentle (not excesive force) left-write movements.

Panagiotis

 

I don't trust it at all, but am experimenting writing, deleting, connecting, disconnecting in my spare time + a horrendous nuisance going into event log all the time.

No heat issue. It's only plugged in when I want to do something.
I did remove it with left-right movements. Smooth and EASY! why didn't I think of it.

Thank you sooo much!

 

You're welcome.
What OS you use? Is the drive usb2 or usb3? Is it connected on an Intel usb3 controller?
I ask because some versions of Intel's usb3 drivers with some usb3 devices caused such hiccups (disconections, low power supply, etc.). Probably your drive is fine and you need to update the drivers.

Panagiotis

 

Thanks again
- Windows 10 1803 bld 17134.648 on Thinkpad T430s (originally windows7). - I use this drive only via windows10. Win7 and LinuxMint might read but never write.
- The drive is USB3, connected to laptop usb3 port.
- If I read device mgr correctly it's using Microsoft's controller but the name is Intel...
Please, see if you can make any sense of my screen shot.
- I was googling about it (now that I know what controller you all mentioned), and apparently Intel no longer has anything for windows10 so Microsoft takes care of it.
- I have usbhub3.inf dated 4/10/18 in system32. It mentions power saving. As I review some scribbled notes, couple of the "bad block" errors happened on a sleepy drive which perhaps turns on too slowly. I need to confirm that. It may just need me to change that power setting. What do you think?

 

You can try to disable "usb selective suspend" . Ususally this power saving setting and the states s3 and s4 cause such problems (the device is put in lower power draw from the OS/bios/uefi but when it tries to exit from that state the machine does not provide enough power to initialize the spinning of the disk again).

Panagiotis

 

Made sense to me. So I went to change, but it turns out that it's already disabled.
The drive does its own thing - goes to sleep after about 3-5 minutes of inactivity.
If I do safe remove/eject while it's alive, no errors.
If I delay and do safe remove while it's asleep, then event 11 - controller error, every single time.
- I wonder if I'm writing to disk and delay safe remove whether it's doing the complete write or something gets corrupted at that point.

My other Convio drives are ok, no such problems. They're different models though. And a WesternDigital is also behaving.

 

What are the drive settings? Is optimized for quick removal enabled?
If you want to prevent it to go to sleep you can use a program like keepalivehd or NoSleepHD to prevent it from entering in sleep mode.

 

I'll have to google about those utilities. You think they apply to just one external drive?
In the meantime:
Not sure I went to the right place to look. I connected the drive.
Through explorer, drive properties, tools, driver details has properties dropdown.
Probably these answer your question:
- Removal policy and default removal policy = 3
- Safe removal required = true
- Power data
current state D0
capabilities 00000019; PD_CAP_D0_supported; PD_CAP_D3_supported; PD_CAP_wake_from_D0_supported.
power state mappings s0 to D0; S1,S2,S3,S4 all to D2; S5 to D3.

Another Convio drive has the same settings but does not get controller error.

 

https://github.com/stsrki/KeepAliveHD
https://archive.codeplex.com/?p=nosleephd

Ok you have the default = optimize for quick removal.
I would not worry about that error at least for an external drive (if it happens only when/after it gets in sleep mode). If it is not spinning (sleep mode) you can remove it without losing data or having ill effects.

edit: also try with another USB cable and check how it behaves... it could be a damaged cable that causes the errors.

Panagiotis

 

Thanks for the links, much appreciated.
Windows is so nasty. Every dialog uses different words. I suppose Removal policy 3 means quick removal. How on earth are we to know that kind of stuff, arrggghhh.

This error ONLY occurs on removing a sleeping drive, Not when I plug it, not when I start a program to access it. Thank you, I'll ignore it.

I did that because you suggested it earlier. I forgot to report, sorry. Swapping cables made no difference.
Thank you so much for your advice and wisdom

I guess at this point I'm still stuck with event 7 bad block and SMART not reporting it, and of course the initial issue why did chkdsk /r mess it up so badly.

EDIT: I just read your links.
I might try NoSleepHD because it allows selecting one drive.
Curiously, it mentions standard sleep after 10 minutes. Mine is much less, more like 3.