Bad sectors

[testcy]

Is there a way to unmark bad sectors and test them again? Is running chkdsk /b the only way?

[porina]

Bad sectors are a sign the disk is on the way out so consider replacing it. If you want to spend time on it, then what I'd do is first empty the disk so there is nothing on it you want to keep. Use the disk manufacturer's tool, or alternatives, to do a FULL surface read. Then a FULL surface write, and finally another FULL surface read again. Why that order? First read will detect any suspect or bad sectors. The write pass will give the drive a chance to remap them with good spares. The last read pass is to check that. If there are still bad sectors at this point, really consider replacing. After that you can reformat and use with care.

[testcy]

3 minutes ago, porina said:

Bad sectors are a sign the disk is on the way out so consider replacing it. If you want to spend time on it, then what I'd do is first empty the disk so there is nothing on it you want to keep. Use the disk manufacturer's tool, or alternatives, to do a FULL surface read. Then a FULL surface write, and finally another FULL surface read again. Why that order? First read will detect any suspect or bad sectors. The write pass will give the drive a chance to remap them with good spares. The last read pass is to check that. If there are still bad sectors at this point, really consider replacing. After that you can reformat and use with care.

RAW value for Reallocated Sectors Count is greater than zero. Chkdsk reports 0 KB in bad sectors. Full surface read reports no problems. The drive is only used for backup, still I would prefer to avoid full surface write, if there is another way.  

[kokosnh]

As poring said ( assuming it’s HDD, if it’s ssd then nand is failing so can’t do anything, if it’s second gen SMR HDD ( with trim ), the you also can't do anything )

first empty the HDD ( well you don’t need to, but full write will delete everything, so copy what you need from it )

Do a screenshot of smart table in crystaldiskinfo

You can use MHDD or Victoria to do the rest.

first a full surface sector delay test, with log ( you need to log all slow sectors, usually grater then 200ms, best to do it twice to exclude NCQ errors ) without remap ( if we do it with remap, then even soft bad blocks will be remapped )

and here we can decide what to do with it. ( if it’s good we can stop here )
then you can do a zeroing ( so full write with zero ) deletes everything.

then again full surface sector delay test, with log.

after that we can go back and remap any remaining bad sectors if needed. 
and last do a screen of smart table and compare the two. 
It can take hours to days, to weeks, depending on the size, and damage. 

[Kilrah]

13 minutes ago, testcy said:

Chkdsk reports 0 KB in bad sectors. Full surface read reports no problems.

You won't see chkdsk report bad sectors or do anything about them with modern drives, it's not how it works anymore. The drive handles bad sectors and remapping internally and just tells you how many were remapped through SMART. 

Full drive write then read may clear some pending ones but usually remapped ones remain.

[testcy]

10 minutes ago, kokosnh said:

As poring said ( assuming it’s HDD, if it’s ssd then nand is failing so can’t do anything, if it’s second gen SMR HDD ( with trim ), the you also can't do anything )

first empty the HDD.

Do a screenshot of smart table in crystaldiskinfo

You can use MHDD or Victoria to do the rest.

first a full surface sector delay test, with log ( you need to log all slow sectors, best to do it twice to exclude NCQ errors )

without remap ( if we do it with remap, then even soft bad blocks will be remapped )
then you do a zeroing ( so full write with zero )

then again full surface sector delay test, with log.

after that we can go back and remap any remaining bad sectors if needed. 
and last do a screen of smart table and compare the two. 
It can take hours to days, to weeks, depending on the size, and damage. 

It's CMR HDD, but I would prefer to test under Windows, if it will take that long. What are the steps to follow if using Victoria? 

[Ryker Robb]

57 minutes ago, testcy said:

Is there a way to unmark bad sectors and test them again? Is running chkdsk /b the only way?

I don't think so, chkdsk only prevents the os from using them, it doesn't fix them. That's why you can undo it with a /b. There probably is a way using pc3000 or similar tool, but that's beyond me.

[Ryker Robb]

3 minutes ago, testcy said:

It's CMR HDD, but I would prefer to test under Windows, if it will take that long. What are the steps to follow if using Victoria? 

Basically a read write read, like others are saying, if you know the specific lba's that are bad, you can just skip to them, and speed up the process, instead of doing the entire drive, but that's up to you.

[testcy]

1 minute ago, Ryker Robb said:

Basically a read write read, like others are saying, if you know the specific lba's that are bad, you can just skip to them, and speed up the process, instead of doing the entire drive, but that's up to you.

Why do I keep getting a Security Warning when opening Victoria, that The publisher could not be verified. Are you sure you want to run this software?

[kokosnh]

19 minutes ago, Kilrah said:

The drive handles bad sectors and remapping internally and just tells you how many were remapped through SMART. 

Not always, some hdd will not remap anything on there own, like some seagate etc… and on others like second gen SMR, you can't even force it…

 

4 minutes ago, testcy said:

Why do I keep getting a Security Warning when opening Victoria, that The publisher could not be verified. Are you sure you want to run this software?

Always used Hiren's BootCD to run it, so don’t know…  

[TomChaai]

There is no point testing them again, what's broken is broken, HDD firmware made sure they are real bad when performing read/write tests. All you do is risking crashing the heads by repeatedly scanning over damaged areas.

[testcy]

On 12/10/2024 at 6:20 PM, kokosnh said:

As poring said ( assuming it’s HDD, if it’s ssd then nand is failing so can’t do anything, if it’s second gen SMR HDD ( with trim ), the you also can't do anything )

first empty the HDD ( well you don’t need to, but full write will delete everything, so copy what you need from it )

Do a screenshot of smart table in crystaldiskinfo

You can use MHDD or Victoria to do the rest.

first a full surface sector delay test, with log ( you need to log all slow sectors, usually grater then 200ms, best to do it twice to exclude NCQ errors ) without remap ( if we do it with remap, then even soft bad blocks will be remapped )

and here we can decide what to do with it. ( if it’s good we can stop here )
then you can do a zeroing ( so full write with zero ) deletes everything.

then again full surface sector delay test, with log.

after that we can go back and remap any remaining bad sectors if needed. 
and last do a screen of smart table and compare the two. 
It can take hours to days, to weeks, depending on the size, and damage. 

In Victoria for Windows, I see only Full surface scan with grid, Full surface scan with graph, Erase All data, Erase and verify data. Where are the log settings?

[kokosnh]

Can you do a screenshot of it. I wonder if something changed… 
 

it should be like this 

for now no remap, only read scan, and log slow sectors, you can change log location in setup. 

 

[testcy]

8 hours ago, kokosnh said:

Can you do a screenshot of it. I wonder if something changed… 
 

it should be like this 

for now no remap, only read scan, and log slow sectors, you can change log location in setup. 

 

The log boxes are missing, but there are log settings under settings. I did a read scan with no remap with no warnings, no errors. RAW value for Reallocated Sectors Count remained the same. I tried simple overwrite in SeaTools, but again nothing changed. Is there anything I can do to unmark bad sectors and test them again?

 

 

[kokosnh]

8 hours ago, testcy said:

The log boxes are missing, but there are log settings under settings. I did a read scan with no remap with no warnings, no errors. RAW value for Reallocated Sectors Count remained the same. I tried simple overwrite in SeaTools, but again nothing changed. Is there anything I can do to unmark and test again bad sectors?

 

 

Ok, wait. 
can you post whole smart table screen here from crystaldiskinfo.

If you did a scan and it’s ok. That means that currently the surface is good, and have no bad sectors. 
We can’t do anything to the sectors already remaped, but it looks like currently the HDD is ok. 
 

[testcy]

9 minutes ago, kokosnh said:

Ok, wait. 
can you post whole smart table screen here from crystaldiskinfo.

If you did a scan and it’s ok. That means that currently the surface is good, and have no bad sectors. 
We can’t do anything to the sectors already remaped, but it looks like currently the HDD is ok. 
 

Since I already did an overwrite is there anything else I could try before using the HDD again?

[Ryker Robb]

8 hours ago, testcy said:

The log boxes are missing, but there are log settings under settings. I did a read scan with no remap with no warnings, no errors. RAW value for Reallocated Sectors Count remained the same. I tried simple overwrite in SeaTools, but again nothing changed. Is there anything I can do to unmark bad sectors and test them again?

 

 

Unmarking reallocated sectors will most likely require specialized equipment (like pc-3000), and involves modifying the drives firmware, you aren't going to be able to do it with any of the free tools you've mentioned. There's not really a reason to do this anyways, if the drive reallocated those sectors, they're 99% likely to be in fact bad.

[kokosnh]

This looks like seagate HDD

and well, the thing is, that’s just to much (784) reallocated sectors to trust it, even if the surface is fine now.
you can use it as second backup HDD. 

[Ryker Robb]

28 minutes ago, testcy said:

Since I already did an overwrite is there anything else I could try before using the HDD again?

Not really, you could try an extended self test, but that's about it.

[testcy]

2 hours ago, kokosnh said:

This looks like seagate HDD

and well, the thing is, that’s just to much (784) reallocated sectors to trust it, even if the surface is fine now.
you can use it as second backup HDD. 

It's a Barracuda 1TB and this is what I am using it for now, but I was thinking to put it in more use. What is the worst case scenario?

2 hours ago, Ryker Robb said:

Not really, you could try an extended self test, but that's about it.

Since I already tried a simple overwrite and nothing changed, could an extended self test make a difference?

[kokosnh]

You can not change the already reallocated sectors, it’s already done…. 

as for barracuda it’s SMR HDD, with defective surface ( looking at smart ), use it only as second backup HDD.

It’s fine now ( as you did surface test ), but when you will try to read your backup in couple months, hundreds of files can be corrupted. It’s better then nothing, but do not trust it with data, or main backup. 

[testcy]

22 minutes ago, kokosnh said:

You can not change the already reallocated sectors, it’s already done…. 

as for barracuda it’s SMR HDD, with defective surface ( looking at smart ), use it only as second backup HDD.

It’s fine now ( as you did surface test ), but when you will try to read your backup in couple months, hundreds of files can be corrupted. It’s better then nothing, but do not trust it with data, or main backup. 

Actually, it's CMR and not SMR.

[kokosnh]

8 minutes ago, testcy said:

Actually, it's CMR and not SMR.

Older ones were CMR, the new ones are SMR.

if it’s older, then yes, it’s CMR. 
we don’t know the model, as you never mentioned it, so… 

[testcy]

2 hours ago, kokosnh said:

Older ones were CMR, the new ones are SMR.

if it’s older, then yes, it’s CMR. 
we don’t know the model, as you never mentioned it, so… 

ST1000DM003

[Ryker Robb]

9 hours ago, testcy said:

It's a Barracuda 1TB and this is what I am using it for now, but I was thinking to put it in more use. What is the worst case scenario?

It suddenly dies, and you lose all data on it. Still not super likely, just something to be aware of.

9 hours ago, testcy said:

Since I already tried a simple overwrite and nothing changed, could an extended self test make a difference?

Most likely not.