Corrupted drive

[sicknero]

Thanks for the comments everyone

Ok .. the exact nature of the problem;

The hdd in question used to contain two folders ... one full of music and the other of audio books.

What happens is that when I play some of the music files, I find that I'm listening to a chapter from an audio book. According to the filename and location I should be hearing a song, but it seems that the actual content of the file has been replaced.

For example, I play "H:\AUDIO\Music\Belly - Discography\Star\01 Someone to Die for.mp3" but what I hear is a chapter from an audio book instead.

The audio book directory I moved to another drive quite some time ago, so the drive in question now contains only music. Or should, at least. As Remah says, at some point the file entries became incorrectly linked.

The presence of duplicated filenames would indicate files that need to be checked.

This is a good suggestion, because a second issue has come to light ... some files have become duplicated without the filenames or paths changing. For example I play tracks 1 and 2 from an album but find that track 2 is actually a duplicate of track 1, although according to the filename it should be track 2.

However, as it is the file content and not the name which I need to check, it will mean running a duplicate finder set to search content but not names ... given the size of the drive and that it is close to being full, this will take hours so I'll leave it to run overnight. It will still be a time saver though compared to checking each song by ear.

I also assumed that all the filenames were the names of recognizable music track. That is normal. But what if there are filenames that have been converted so they aren't. Have you checked all the filenames in all your music folders to see that they are all as expected?

Yes ... the filenames and apparent directory structure is are all exactly as they should be.

In forensics, the number 1 essential step is to image the drive and work on that instead of modifying the drive itself. That means you'll have a backup and a workspace that won't corrupt any further.

What I'm doing now is checking each folder in turn using a media player set to intro scan ... if the folder is ok then I'm copying it across to a new drive, so gradually building up a new directory of uncorrupted files. It's taking some time as given the size of the collection, there are inevitably albums with which I am not sufficiently familiar so as to instantly know if a song is misnamed .....

I am pretty much resigned to the fact that there's no easier way to do it really, especially given the amount of time that has passed and the amount of changes that have been made to the drive since the corruption first happened. For example I don't think there's much hope of recovering deleted files as there has been so much over-writing in the meantime.

[sicknero]

... afterthought ...

Given that on the whole audio book chapters are substantially larger files compared to songs, then sorting files by size might be one easier route to finding out which files have been misnamed.

[Remah]

Now that many users have large quantities of data, the need for comprehensive data protection becomes more important. I regularly hear of users finding errors in their many TB of photos, music, videos, etc. They usually only find the problems when they get an error transferring the files to a new (and usually larger) drive.

For several years, my preference has been to use ZFS to ensure data integrity. ZFS is available free with FreeBSD and free products based on it, e.g. FreeNAS. Regular "scrubbing" (checking files and data) even picks up and repairs silent errors which become more common with larger volumes of data.

ZFS is not the only free file system to provide advanced data protection but it does have the longest history providing it to users like us. Most other solutions require the use of multiple products to get the breadth of features. However Linux does have a comparable solution with BtrFS. BtrFS is more actively developed and better supported than the Linux version of ZFS. It also performs many of the same functions as ZFS, i.e. a combination file-system, logical volume manager, and software RAID subsystem.

The article The Cult of ZFS explains this in more detail.

[kendall.a]

OMG...Remah, is there anything you don't know something about?! You are simply amazing with your knowledge-base!

Half of what you just wrote, I don't understand, but it sounds really cool.

Thank you for being here and for sharing your knowledge with us. I, for one, intend to look at this program and read your link. Interesting stuff....

[sicknero]

Indeed, thank you. The idea of adopting a new file system is quite intimidating and raises a lot of questions ... I shall follow up the links and try to digest all the information : )

[Remah]

It's worth being aware of such products and features which are often built into NAS products you buy at a computer shop.

If you are a Windows user then Microsoft's new file system ReFS provides many of the same features as ZFS abd BtrFS. ReFS will eventually become more mainstream, maybe with Windows 10. It is available with Windows Server 2012 and 64-bit Windows 8.1 but is far from a mature product and you'd really want to know what you are doing before using it.

ReFS uses Storage Spaces, Microsofts new logical volume manager which also provides software RAID. It is available for NTFS so I hear of many people using it but again it is better to have some technical understanding of it.

So comparing the Windows products with ZFS, there are three main functions that ZFS provides which you could largely match on Windows using ReFS and Storage Spaces. It's just that ZFS is still more reliable:

1. File system provides the means to organize data into files, folders, etc. Examples are ReFS, NTFS, FAT
2. Logical volume manager allow groups of disks and other storage devices to be organized almost any way that you like so for example, a logical drive can be spread across many physical drives and can expand and contract beyond the size of any one physical drive. Examples are Windows Logical Disk Manager which you've probably never heard of or used which is now being replaced by the more advanced Storage Spaces.
3. Software RAID manager allows RAID data redundancy to be provided without any special hardware. Software RAID is built into Storage Spaces for example.

[sicknero]

Sorry if this is a little O/T but this thread has brought a related issue to mind for me on the subject of corrupted drives.

I've read at times in various forums, that the Windows diskchecker can sometimes do more harm than good ... I'm just wondering what thoughts if any people have on this..?

[Remah]

Originally Posted by sicknero

Sorry if this is a little O/T but this thread has brought a related issue to mind for me on the subject of corrupted drives.

I've read at times in various forums, that the Windows diskchecker can sometimes do more harm than good ... I'm just wondering what thoughts if any people have on this..?

I've not had any of the specific problems that I've read about many over the years. But I have noticed the following issues.

1. It is used when other products are better suited to the task at hand
Part of the problem is that Windows check disk is used for problems that it can't really help with. Some problems can even cause it to crash.

It is not a complete hard disk check. Rather it is a logical disk/partition check that is little or no use with partition management, boot records, and physical disks/drives.

Other products do a better job because they can recover damaged structures and data more effectively. Plus they can check the physical hard disk and recover and repair it more fully.

These two categories have some better products:
Best Free Data Recovery and Undelete utility
Best Free Hard Drive Health Monitoring and Diagnostic Programs


2. Checking can accelerate the occurrence of problems
The act of comprehensively checking an entire drive can accelerate the occurrence and discovery of problems that would not have either occurred or been discovered until a later date. This becomes more obvious the more finite the life of the media.

Specific tests also increase the stress on particular components so the failure of that component is more likely. For example, a random disk read test is more likely to induce a failure than a sequential read test which doesn't move the read heads as much.

Something similar happens with memory checks too.

[sicknero]

Thanks again Remah.

I was asking because my habit is always to run chkdsk in the event of drives not being properly dismounted ... e.g. a system freeze forcing me to reboot using the power switch. I was wondering if that might have been a contributing factor to my hdd problems.

[Burn-IT]

Running CHKDSK is never going to cause problems with any disk that didn't already have problems. That is a complete lie. CHKDSK will do no more head movements than any other tool.
Checking a disk is NOT going to stress any components. It it always going to be better to run CHKDSK if you suspect bad dismounts.

You will try telling us next that SCANDISK was a bad tool for FAT32 disks or that the old Norton Utilities were bad.