CRC Error ?

[Deist]

Hello there!

I have been enjoying the high efficiency of HDSentinel for more than one year now, and feel much safer about my datas thanks to it!
Unfortunately, one of my 1Tib HD is in a USB enclosure that according to HDS is not sending any SMART information, and worse thing is the enclosure is sealed shut so that I can't remove the HD and put it in new compatible enclosure in order to check it.
I have had and used (quite a lot!) this HD for about 1 year now, and this morning I got my first "Cyclic Redundancy" error while attempting to compress (with WinRAR) a bunch of WAV files! One of the files (which was previously extracted without problem from a ZIP archive) is now corrupted and unusable due to making this cyclic redundancy error every time it is accessed.

My only question about this is: does this means my 1 year old external HD is in serious danger and about to fail completely?
Or may cyclic redundancy errors appear on a HD without necessarily being the symptom of something way worse about the health of the HD?


I thank you for your time reading this, and am really looking forward to your answer.
Regards,
-- DeisT

[hdsentinel]

Hi Deist,

CRC errors and data transfer errors in most cases are not really hard disk issues.

These may be related to the data communication channel: the USB cable or connection in this case.
I recommend to connect the drive directly to a rear USB connector with the shortest possible USB cable and avoid using the front USB and avoid and USB hubs/extenders when connecting this drive.

Insufficient power is also a common issue for data transfer (CRC) errors. I suspect it is a 3.5" drive so it may not be affected, but many 2.5" USB drives are suffering from this problem (except if two USB slots are used with an Y USB cable).

Overclocking, memory problem, overheat (not only the hard disk but also CPU/chipset/RAM overheat) are also common causes of data transfer errors.

Anyway, the best would be to disassemble and connect the drive directly to the motherboard or using in an other enclosure which allows disk monitoring by Hard Disk Sentinel. This would immediately show if there is a serious error with the drive or if the drive itself logged any data transfer / communication errors at the "Ultra ATA CRC error count" attribute. But as you wrote it is not possible ....

I also recommend to use the tests at Disk menu Surface Test. These tests can diagnose the disk even if it is used in this enclosure which does not allow detection of further disk information. Some of these tests are destructive so be careful about the stored data.

[Deist]

Many thanks for your comprehensive answer! I take good notice of your comments and advices. I will indeed find the shortest possible USB cable to use with this enclosure, and will use the surface test on the blocks where the corrupted file is.

Incidentally I should have precised that I'm using this external 3.5'' HD with a laptop, as I don't have a desktop computer (at the moment anyway), no overclocking, no USB hub, and both computer and external HD are well ventilated to prevent overheating.

I also realise I may not have explained my problem correctly talking about a "CRC error", as I'm not talking about a file being merely corrupted here: it is not a usual CRC __checksum__ error that makes the file unusable. I am in fact talking about a file which can't even be compressed and put into an archive! When trying to do this, the first half of the file is being processed correctly, at which point when reaching like 60% in the file the OS itself reports an error while accessing it, and WinRAR reports a "Read/Data error in the file (cyclic redundancy control)" which not only renders the file unusable (like a common CRC checksum error would), but actually renders it unaccessable too (except for deletion)! (As opposed to a more common CRC checksum corrupted file which would still be readable by the OS and could be compressed, or processed by applications).

In my past experience, I have seen this kind of error happen more and more with a couple of dying HDs, until the drive turns finally completely unusable, so I am very worried to see that problem happen on this relatively new Hard-drive being only about 1 year old!

What made me suspicious about a problem related to the hard disk itself in the 1st place is the fact that the reading error is not random and occurs each and every time with the one same file, and rigorously at the same precise part in the file (when trying to compress the file with WinRAR or 7z, the process will get stuck and report the access error always at the same precise spot). Thus far, it is the only file which raises this problem and I didn't have trouble with any other. Also, this is the very first time such trouble arises after using the drive for 1 year without encountering any.

So allow me to ask you again what your opinion is regarding that error: is it likely to be a symptom of deteriorating health?


I thank you once again for taking the time to read this.
-- DeisT

[hdsentinel]

I agree, this is a different scenario.
I thought the issue was that the data is readable, just the CRC value did not match (so the data was corrupted during transfer).

Now I see it is really related that one (or more) particular blocks which cannot be accessed at all.
This is most likely caused by a weak sector: a sector which seems to be not reliable and may be reallocated in the later usage.

In this case, the Disk menu Surface test options can help:

1) First of all, I'd recommend a READ test to verify which parts of the disk surface is affected and how many such sectors found. This test is safe for the stored data.

2) To regenerate the disk (fix the issues by forcing reallocation of sectors) without clearing the current data, you can use the "Read + write + read test (refresh data area)".

3) In case of too much or serious problems, the "Reinitialise disk surface" would be the best option to diagnose and fix these sectors if possible. Be careful as this test clears the stored data.

The best would be if the drive can be monitored (in a better a disk enclosure) during the tests for any changes in disk parameters. Anyway, in the current environment the tests can help to examine the disk usability and performance during the test to verify if some parts of the disk is hardly or slowly accessible - which may be real signs of disk health decrease.

[Deist]

Thank you so much for the valuable and extensive advices!!!

Following your suggestions I did the Read surface test with HDSentinel (left it to work all night as a full read for a 1 Tib HD took a good 9 hours long). The final result of this is that my HD has 2 contiguous bad (red) blocks (#7544 and #7545), otherwise all other blocks are reported good (green). I then proceeded to Read+Write+Read test as you adviced to try and regenerate the sectors contents, but it proved unsuccessful to reallocate so the blocks remain bad...

Now I must say I don't care much about the possible data loss (so long as I can figure what files are actually affected in order to possibly download them again)*, considering we're merely talking about no more than 2 contiguous blocks. What bothers me way more is the fear and possibility that these 2 bad blocks are in fact indicative of a deteriorating health (ie. progressing) and soon more sectors/blocks may turn damaged or even bad as well!

What is your opinion? Is this likely? Or is it possible these were bad from the start when I bought the HD and that the status of the drive is stable?


Thank you again for your time and your comprehensive answers about my little problem!
-- DeisT
*(As a side note, would you know of a good software/method to check what FILES or part of files are located in these 2 blocks ? -- as checking the sectors contents is obviously of no help to figure this out)

[hdsentinel]

I see. If at the moment you do not care about the possible data loss but want to repair the hard disk, the quickest method would be the following:

1) Select Disk menu Surface Test option
2) in the window select the drive and the "Reinitialize disk surface" test
3) click on Configuration tab and locate "Limit testing to specific data blocks" and enter the first (7544) and last (7545) blocks you noticed that there were errors (which were red)
4) click on Start test

This way only those specific blocks will be tested and repaired. I suspect it should not take much time but it depends on the severity of the error. Please note that this during this test the drive may not be accessible and it is possible that the software or Windows does not respond quickly. This is normal and please do not power on/off the system or the drive.
Please note that the blocks will be overwritten (cleared) so any file(s) at that position may become corrupted.

If this step is completed, maybe you can try the Read test again, limited to the bad spots (or maybe testing neighbour blocks also, so lets say do a read test from block 7500 to 7600). I suspect (and hope) then it will be completed without errors. If not, maybe you can repeat steps 1-4 but before step 4) you may increase the "Surface reinitialisation level" (Configuration tab, lower right area) from 3 to 10 or even more.

I agree that (even if the sectors will be fixed) you'll worry about the data. Without the ability to constant monitoring (due to the restrictions of the USB enclosure), I recommend to perform a such read scan periodically to check if things are still correct.

The issues may not be related to physical damages. Do you remember any power loss, unwanted disconnection / restart of the system or the drive? These may also cause some bad spots.

At the moment I do not know any software or solution which can display the file(s) damaged due to the bad spots. This feature is under development (with high priority) so a later version of Hard Disk Sentinel will surely be able to identify the possible damaged files on the disk map.

[Deist]

Once again, thank you so much for your time and your answer which proves considerably helpful with my problem!

I agree that (even if the sectors will be fixed) you'll worry about the data. Without the ability to constant monitoring (due to the restrictions of the USB enclosure), I recommend to perform a such read scan periodically to check if things are still correct.

Thank you for the advice, yes. Being unable to monitor the drive, performing a read scan again from time to time was indeed what I planned to do to ensure things are not worsening...

Incidentally I have a last question if that's ok: The Read test being quite a time-consuming one (about 9 hours with my drive), can you tell me to what extend performing it itself impacts on the lifetime of the drive?

The issues may not be related to physical damages. Do you remember any power loss, unwanted disconnection / restart of the system or the drive? These may also cause some bad spots.

I don't really remember any of these happening, but in a one year time, it may well have been, I can't tell... All I am positive about is I always make sure the drive is not working before unplugging it or shutting down the OS myself, of course.

At the moment I do not know any software or solution which can display the file(s) damaged due to the bad spots. This feature is under development (with high priority) so a later version of Hard Disk Sentinel will surely be able to identify the possible damaged files on the disk map.

This is wonderful news! And proof if need be that, although HDS already IS the leading software about HD monitoring (and as far as I know I've tested most of them, if not all (10+), you people developing it still are concerned about finding further possible ways to improve it yet again!

A software worth buying if I ever saw one! Thanks again for dedication and efficiency! (^_^) You sure make my (digital) life easier with helping securing datas (at least until the distant future day when actual long-life archiving disks become affordable on the mass-market, hehe ^_-).

[hdsentinel]

Incidentally I have a last question if that's ok: The Read test being quite a time-consuming one (about 9 hours with my drive), can you tell me to what extend performing it itself impacts on the lifetime of the drive?

The test has no negative impacts on the drive (does not harm the drive). It may even help the drive as during the complete surface analysis it can find any possible issues which would otherwise remain un-noticeable until they can lead to data loss.
This is true if the drive (enclosure) has proper cooling, otherwise it's temperature may become so high just like during any longer disk operation (eg. copy/move huge amount of files, antivirus scan, defragmentation, etc.)

I agree it is really time-consuming, because of the USB connection which limits the transfer speed to around 30-35 MBytes / sec. If a such drive would be connected directly to a SATA channel (or through an eSata interface) the total time should be less than 3 hours.

I don't really remember any of these happening, but in a one year time, it may well have been, I can't tell... All I am positive about is I always make sure the drive is not working before unplugging it or shutting down the OS myself, of course.

I see. This way you do much more than average users to keep the drive (and stored data) safe
Only the recent days / weeks are important (not the whole lifetime of the disk): the days elapsed since the appropriate file(s) are stored on the disk.

This is wonderful news! And proof if need be that, although HDS already IS the leading software about HD monitoring (and as far as I know I've tested most of them, if not all (10+), you people developing it still are concerned about finding further possible ways to improve it yet again!

A software worth buying if I ever saw one! Thanks again for dedication and efficiency! (^_^) You sure make my (digital) life easier with helping securing datas (at least until the distant future day when actual long-life archiving disks become affordable on the mass-market, hehe ^_-).

Thank you very much for your kind words! I'm so glad that Hard Disk Sentinel could help We try to make it as useful as possible and improve it further.

[alan-0000]

Defraggler from Piriform may be of immediate help.

It has just taken 3 seconds to Analyze My Secondary HDD Partition E:\ 466 GB capacity with 352 GB Used Space.
It will obviously take much longer to defragment

After 3 seconds Analyze I see a representation of fragmentation as colored squares in an array 60 blocks high by about 160 (I ran out of fingers) squares wide.
I can click on any square and immediately see the name(s) of files that use that space.
There are about 10,000 squares representing 466 GB, so each square holds about 0.01% of the partition capacity.
N.B. 10,000 squares with the central region maximized. By reducing the width and height of the display the squares retain their size and thus you can increase the percentage of the partition to give a "nice" integer ratio of squares to blocks

If you can relate your defective blocks to a percentage location and hence a specific square(s),
then clicking on the square(s) will tell you the file(s) that are there.

[firionicable]

Can HDSentinel can be used for both HDD and SSD?

[hdsentinel]

Yes of course, both hard disk drives and SSDs are supported and their complete status is detected and reported.