Bad Sectors on Hard Drive

[easysurfer]

Spent yesterday trying backup software only to discover that my system hard drive has bad sectors. I initially thought the bad sectors were on other drives, so looks like I'll be spending the day scanning, scanning and scanning some more.

Somewhere between November 2025 (a last good drive image) and now sectors got bad.

 

[target2019]

Instead of scanning, scanning and scanning, I would retrieve the important files from that drive.

 

[easysurfer]

I do want to let the scan finish (in about 2 hours) to see how many bad sectors are found. I ordered another drive that should arrive in a few days as a replacement.

 

[easysurfer]

Not related to the tech issue but I'm a little suspicious. I made an Amazon order for the replacement hard drive. Not scheduled to arrive until a few days from 3rd party seller. I realized that I have a drive in other computer that I could use as a replacement .

So I issue a cancellation request. I get an auto reply of maybe can't cancel because seller was notified. Then about 2 minutes after this, I get an email showing item was shipped via USPS. The skeptic in my thinks that as soon as the seller was notified of the cancel request he probably started a USPS sent just so he could say the item has been sent and can't be cancelled. You think?

 

[target2019]

Well, it's Amazon. Everybody Lies...

 

[Sojourner]

I do want to let the scan finish (in about 2 hours) to see how many bad sectors are found. I ordered another drive that should arrive in a few days as a replacement.

I hope you ordered an SSD, ideally an NVMe. These days, SSDs are so fast, reliable, quiet, and energy-efficient that I would always use them as primary storage devices for any important data. The only strong use cases I see for old-school HDD technology are bulk cold storage/archiving and for NAS media streaming (due to the much lower dollar/TB pricing).

 

[ShokWaveRider]

Hard Drive, what is that?

Just kidding, I agree with Sojourner, replace with a SSD and do not look back. SSDs are typically only used for security systems and media players and they are a special type specifically designed for those purposes.

 

[Al18]

From the picture in the original post, that's a 1 TB Toshiba SATA drive that failing. A new 1 TB SATA from WD will run you about $75, while a new 1 TB SATA SSD will run you about $150. SSD are faster, but not a lot faster if you're connected to a SATA III 6 GB/s port. Just guessing, but the computer might be 5-10 years old and not have a faster hard drive interface.

 

[easysurfer]

Scan finished showing 3 bad sectors on my hard drive.

 

[easysurfer]

I'm not the most adaptive to change .

 

[Qs Laptop]

Scan finished showing 3 bad sectors on my hard drive. View attachment 63069

Isn't there a way to quarantine those bad sectors and continue to use the hard drive?

Or is this the tip of the iceberg and more bad sectors would follow?

I agree with going with an SSD instead of a mechanical drive, if you decide to replace it.

 

[Al18]

I would continue to use a drive with just 3 bad sectors - that's insignificant.

 

[easysurfer]

I would continue to use a drive with just 3 bad sectors - that's insignificant.

The drive with the 3 bad sectors don't play nice with a couple of backup software that I'm using. Otherwise, I'd probably use the just ignore until things get worse approach.

 

[Ian S]

From the picture in the original post, that's a 1 TB Toshiba SATA drive that failing. A new 1 TB SATA from WD will run you about $75, while a new 1 TB SATA SSD will run you about $150. SSD are faster, but not a lot faster if you're connected to a SATA III 6 GB/s port. Just guessing, but the computer might be 5-10 years old and not have a faster hard drive interface.

Egad, I'm glad I got the NVME SSD's I needed a year and a half ago. I paid $102 for a 2 TB off-brand one. Amazon says it's now $299! Thanks AI

 

[easysurfer]

Update:

I think this is gonna work .

After several days of trying different things like running a chkdsk and marking the bad sectors on the bad drive and trying to then clone with macrium reflect which didn't work for me (had a permission error), I decided to cough over some money and go in blind faith for the marketing pitch of Aomei Backupper (there is also a just clone version) and bought a license.

Using Aomei performed a clone and from my scan in progress, looks like the software did indeed avoid bad sectors .

 

[easysurfer]

The hard drive scan to check for bad sectors is almost done. So far so good, no bad sectors!

Probably overkill, but now I have 4 programs to use for my computer back up and restoring needs.

1. Aomei Backupper Pro - (paid version - I have a lifetime upgrade) - cloning. Can even clone a drive that has bad sectors. Backupper also has functionaliy like my old paid Macrium Reflect v6 so I can use Aomei in the future if things no longer work with Macrium

2. Macrium Reflect v6 - I both this paid version a long time ago. My go to for scheduled, automated incremental backups at night. Happy sectors = happy Macrium. I will continue to use this until it doesn't work anymore with my computer

3. Gnome Disk (free)- I use this when I want to make a full image of hard drive. Can do also with Macrium. But hey, this is linux .

4. MultiDrive (free)- I've never heard of this program until the recent experience with the bad sectors on hard drive. This can do copy even if there are bad sectors and save file as a RAW file. Handy hard drive is messed up. Make copy to have a backup before messing around trying to fix bad drive.

 

[SecondCor521]

I used to work in R&D for a leading SSD manufacturer. (As a minor historical note, the first production firmware for the first SATA III SSD first existed on my hard drive since I was the one who pushed the "build the firmware" button on that program.)

I've been gone from that company for 10 years now, but when I worked there we used Acronis for cloning. There were some Windows security measures that it seemed to handle well.

Bad sectors have been a thing for decades now. All hard disk cloning tools should handle them just fine.

chkdsk -f should find and mark the sectors bad and allow the drive to function for a while longer, possibly a lot longer. Three sectors doesn't seem like that big of a deal, but I'd probably replace the drive anyway like you're doing.

...

One possible interesting note:

The way SSD flash memory works, there is a "bathtub curve": sectors go bad initially at a low but still relatively high rate, then don't go bad for a long time, then go bad as the sectors get to end of life.

Because customers wouldn't be happy to see those initial bad blocks, the company I worked for would run newly manufactured drives for a day or two to expose the initial bad blocks. They'd factory mark them in a way that the drive looked to the customer like it had zero bad blocks. There were extra / spare blocks on the drive that were mapped in to cover for the bad blocks so that the customer saw the correct amount of total blocks on the drive.

This was just inherent in the way the flash memory works, so I'm pretty sure all SSD manufacturers do essentially the same thing on their production lines. This may have also been done on regular hard drives for similar reasons; if so it wouldn't surprise me at all.

Pretty much every new SSD had a handful of these initial bad blocks. Depending on drive size, somewhere between 1 and 10 IIRC. After that initial burn in period, it was really quite rare to have what was called a "grown bad block". Even so, grown bad blocks (like the three mentioned here) were just marked and set aside and other good blocks were mapped in. SSDs can go ridiculously long by this remapping because they have several percentage points more sectors than the standard nominally requires. (It also takes a lot to make an SSD sector go bad; there's error correcting codes that can correct for a surprisingly large number of bad bits.)

 

[Al18]

SecondCor521, thanks for the explanation.

 

[easysurfer]

The hard drive scan to check for bad sectors is almost done. So far so good, no bad sectors!

Probably overkill, but now I have 4 programs to use for my computer back up and restoring needs.

1. Aomei Backupper Pro - (paid version - I have a lifetime upgrade) - cloning. Can even clone a drive that has bad sectors. Backupper also has functionaliy like my old paid Macrium Reflect v6 so I can use Aomei in the future if things no longer work with Macrium

2. Macrium Reflect v6 - I both this paid version a long time ago. My go to for scheduled, automated incremental backups at night. Happy sectors = happy Macrium. I will continue to use this until it doesn't work anymore with my computer

3. Gnome Disk (free)- I use this when I want to make a full image of hard drive. Can do also with Macrium. But hey, this is linux .

4. MultiDrive (free)- I've never heard of this program until the recent experience with the bad sectors on hard drive. This can do copy even if there are bad sectors and save file as a RAW file. Handy hard drive is messed up. Make copy to have a backup before messing around trying to fix bad drive.

Looks like I may have to consider switching over to Aomei Backupper Pro sooner than later. Recent Microsoft Security update causes trouble in paradise . Now wonder my normal Macrium backup is taking so long and encountered an error message.

While it is always good practise to install security updates as soon as possible, sometimes there are unforeseen consequences. Unfortunately, Microsoft has something of a track record in the update failure stakes, from disappearing passwords to broken VPN connections. The latest security update issue stems from the installation of the April cumulative Windows 11 security update, reference KB5083769, which has now been reported as causing backup applications from multiple vendors to fail due to a service timeout issue.

I reached out to Microsoft for a statement and a spokesperson proviued the following: “In the April 2026 Windows security update, we added known vulnerable kernel driver psmounterex.sys to the Vulnerable Driver Blocklist. Backup applications that rely on this driver may experience failures when attempting to mount or manage disk images. We do not recommend uninstalling or pausing this update. Customers with an impacted driver should install the latest application versions and validate against the driver blocklist to remain protected.

Microsoft Update Warning—Windows 11 Security Fix Breaks Backups

Windows 11 users are warned that backups are broken across multiple third-party applications after another Microsoft security update brings unexpected consequences.

www.forbes.com

 

[Sunset]

I would use an SSD for the main drive, I do use a spinning old drive for storage of pictures and music, etc.

The difference in copying files is incredible, along with scanning the drives, etc. SSD's are incredible.

So glad I bought a couple of 2TB SSD's with heat sinks (which reduce temp by 5 degrees, according to my readings) as the price has doubled and is no longer $114.

 

[fognostic]

Check the SMART data for the drive. In particular look at the following SMART ID's:

5 reallocated sector count (you have 3 bad sectors - if this starts climbing, replace ASAP);
187 reportable uncorrectable errors (if reported and >0 replace ASAP;
188 command time out (if reported and into the hundreds, replace ASAP);
197 current pending sector count (>0, replace the HDD);
198 offline uncorrectable (>0 replace the HDD ASAP);
200 multi-zone error rate (if this rises along with 5, replace ASAP).

How old is the drive? Look at SMART ID 9:
9 power on hours (if this is > ~53,000 hours - ~6 years - you are flirting with the rising end of the bath tub curve and it's time to think about replacing the drive)

Since you are running GNOME Disks on linux, you can run SMART tests and get the SMART data connecting the drive to your linux machine (you might need to also install the smartmontools package). Since the drive is installed a Windoze machine, you can at least get the SMART data off the drive from the free CrystalDiskInfo utility without having to swap it over to another machine, but this won't run SMART tests for you like GNOME Disks.

Overall, it seems to me you might be playing with fire trying to keep a dying drive alive if you value the data on it.

 

[easysurfer]

Check the SMART data for the drive. In particular look at the following SMART ID's:

5 reallocated sector count (you have 3 bad sectors - if this starts climbing, replace ASAP);
187 reportable uncorrectable errors (if reported and >0 replace ASAP;
188 command time out (if reported and into the hundreds, replace ASAP);
197 current pending sector count (>0, replace the HDD);
198 offline uncorrectable (>0 replace the HDD ASAP);
200 multi-zone error rate (if this rises along with 5, replace ASAP).

How old is the drive? Look at SMART ID 9:
9 power on hours (if this is > ~53,000 hours - ~6 years - you are flirting with the rising end of the bath tub curve and it's time to think about replacing the drive)

Since you are running GNOME Disks on linux, you can run SMART tests and get the SMART data connecting the drive to your linux machine (you might need to also install the smartmontools package). Since the drive is installed a Windoze machine, you can at least get the SMART data off the drive from the free CrystalDiskInfo utility without having to swap it over to another machine, but this won't run SMART tests for you like GNOME Disks.

Overall, it seems to me you might be playing with fire trying to keep a dying drive alive if you value the data on it.

The drive is no longer having issues.

My latest post is about how Windows update borked backup software like Macrium, Acronis. I probably should have started another thread to avoid the confusion.

More on that here:

Microsoft’s April 2026 security update was supposed to protect Windows users from a serious threat: a buffer overflow flaw (CVE-2023-43896) in the psmounterex.sys driver, which could allow hackers to escalate privileges or run malicious code. In response, Microsoft added the driver to its Vulnerable Driver Blocklist—a move that, while critical for security, has left a trail of collateral damage.

Backup software from major vendors—Macrium Reflect, Acronis Cyber Protect Cloud, UrBackup Server, and NinjaOne Backup—relied on the now-blocked driver for mounting backup images and creating VSS (Volume Shadow Copy Service) snapshots. With the driver disabled, users can still create full image backups, but browsing or restoring those backups fails. The error manifests as a VSS timeout or the cryptic code VSS_E_BAD_STATE, and deeper investigation in Event Viewer reveals Code Integrity errors (Event ID 3077) directly referencing the blocked driver.

Windows Security Update Leaves Backup Tools in Limbo: Critical Driver Blocked, Data Restores Fail

April 2026’s Windows update disables a key driver to fix a major vulnerability—crippling backup software and putting system recoveries at risk.

netcrook.com

 

[jim584672]

Some Linux file systems (Btrfs and Zfs) use self correcting algos to fix bit rot.

 

[GenXguy]

If a hard drive is showing bad sectors, then there were already a bunch of bad sectors that weren't reported in the reserved spare sector pool. So those visible bad sectors are indicating imminent drive failure.

 

[SecondCor521]

Some Linux file systems (Btrfs and Zfs) use self correcting algos to fix bit rot.

That's what the SSDs are doing in hardware and firmware under the hood too; they just don't take up user data space to do so.

The SSD UBER is ridiculously good for consumer use - I looked at one the other day that had an UBER of less than 1 sector in 10^17 reads. Where one sector probably means just a bit or two that exceed the correction power. How long does it take for a consumer drive to read 10 ^ 17 sectors? Too lazy to do the math but it's probably decades for the average user.

There are also a lot of other mechanisms going on under the hood - they'll move data if they see a hardware sector going bad; some can do RAID at the drive level, there is wear leveling which helps preemptively, and probably others I've since forgotten.