Can corrupted RAM damage files

vbimport

#1

I was wondering if corrupted RAM can damage files on my hard drive such as my music, photos, documents? I recently had corrupted RAM, which i didn’t replace for a while. While the corrupted RAM was in my computer, I transferred many files to my hard drive and also had many files on it. I have now replaced that RAM and reinstalled the OS, but would it have damaged my files in any way?


#2

Files on a harddrive which aren’t updated would be mostly safe, but any files that were written while you had corrupted RAM could be bad, and you wouldn’t even necessarily find the corrupted files by doing a filesystem check, although you might find some of them way.

I hope you have a backup of the files you copied to that harddrive.

    • Moved thread from Flash Memory forum to Hard Drive forum * *

#3

What do you mean by not updated? Is it a high possibility that those files written while i had corrupted RAM are damaged? Oh and luckily I still have backups of some of my files.


#4

If there were files already written on your harddrive before the corrupt RAM problem started, and you didn’t modify those files while you had corrupt RAM in your pc, then those files are probably safe.

They are not guaranteed to be safe, however, because with corrupt RAM there’s a small risk that your pc would write to the wrong location on your harddrive thus overwriting part of another file.

Oh and luckily I still have backups of some of my files.
If I were you, I’d check the drive for errors (e.g. Properties => Tools => Error-checking on the drive icon in Explorer). And then I’d restore those files from backup that are supposedly not updated on your harddrive - on the off chance that they have been corrupted.

And then I would check any new or updated files to see if they have the right content, because the Windows Erro-checking only checks that the structure is correct - it cannot check that the files have the right content. This might be more files than you have time to check, and in that case I’d check all the most important ones and only check some random subset of the remaining files.


#5

ok. Thanks for your help


#6

why do you think your ram is bad ?

here is free sftwr to test your ram. If you do not have a floppy they have an iso you can burn to a cd.
http://www.memtest86.com/


#7

I had already used memtest to diagnose my first RAM which was corrupt. Thanks anyway.


#8

If structures like the NTFS MFT end up passing through the bad RAM, a serious mess can be made.

If it errored only once, in several test passes, then it’s possible something else (maybe power noise) caused it to glitch, though throwing even a single error means it is not entirely trustworthy. Also, if overclocked, slacken it.

Mind you, it’s usually said that it takes MANY test passes to prove RAM good, and only one fail to prove it bad. If it can repeat the same test without error, it shows the error was transient, the question is, HOW transient, as it is possible to have spurious errors, that was the original reason for parity on memory - and if you got one parity error, you’d see if it repeated, rather than condemning the RAM immediately


#9

Could damaged files, viruses, faulty OS, intensive memory use etc. corrupt RAM in anyway. I was wondering if the possibly damaged files caused by my previous bad RAM can corrupt or damage my new RAM, or if any other factors listed above can also corrupt it?


#10

Any help please?


#11

In a word, no. It is true though, that any of the potential system issues that might cause unstable RAM could have the same effect on new RAM. Such as low voltage, incorrect timings, etc.


#12

thanks for your help


#13

Defective RAM will almost always (like 99.9%) of the time NOT corrupt any locally stored files. Even if they are being copied back and forth frequently. This is the whole reason for CRC. If the original file, and new file do not have the same HASH value, the file is considered corrupted. When you say your RAM is corrupt I assume you mean defective. Defective meaning that what you write into memory, is not what you get back out. If you copied all your files, and you didn’t get any errors while you are doing it. It is very likely you will not run into a corruption problem. If one binary bit is offset, or out of sequence, the new file will FAIL the CRC check. Luckily smart people have thought of this for servers, where if CRC didnt exist, a bad RAM module could potentially rewrite millions of important files in a matter of minutes, before the system hangs.

Other non related problems, like viruses, etc. Could def. pose a problem. RAM is also volatile, meaning that if the system see “bad memory references” it will try not to address them. This is usually futile.

-Brian


#14

CRC checks on a home PC are only done on file reads and writes to verify the data passed over the cable correctly, they are not done internally on the PC in windows. This means that on a file read the CRC is calculated and used to check if the file was recieved correctly, the file is stored in RAM (some RAM uses CRC checking but 99% of home PC’s dont have this as its expensive). This storing in RAM corrupts it because Windows does not use CRC checking of memory cache. The file is then re-written to the new location, a new CRC value is calculated and the transfer over the cable checked again (a seperate CRC value is also used on the HDD disc itself to check for storage errors and recovery, this CRC value can again be different from the cable generated one if there is faulty ram on the HDD.)

I have had lots of instances where bad ram has corrupted file transfers. Sometimes people are lucky and the corruption is never noticed or it is very infrequently used files, sometimes the O/S is trashed because a defrag had been done and most of the files moved were corrupt.

Permanent data corruption cannot happen on files that are just read because they are never changed, new files written with bad memory could be damaged because no checking is carried out on the file that is cached by windows to be written.

CRC is never used through the whole data transfer chain unless you physically do it yourself. Each step of the chain has CRC checking but the value of that CRC is never passed on to the next stage.


#15

I have transferred 1000’s of GB’s of data from broken drives. And I have never gotten a corrupted file unless a CRC error is thrown (Yes I have to check them all). I am a PC Tech, and have done many backups for working and semi working drives.

Can you provide documentation on RAM cache not being cecked via CRC? I am under the distinct impression that only one CRC value is calculated for any given file, whether it is being read or written to RAM/Harddisk.

-Brian


#16

I had a bad experience with faulty RAM module. I already had installed 512 megs of RAM and decided to upgrade. So I added another module of 1024 megs. Bad one. I got fatal NT bluescreen after a few hours. Then I rebooted but overnight my server application had crashed. The operating system could work in these conditions for days if left idle, though. I theorize that it’s because the primary lower memory was good. I did not know enough about hardware then and assumed that the mem modules were not compatible with each other. But after testing the new ram, it turned out to be bad.

All files I had copied while the faulty ram had been installed had errors once in about every 50 megs. Now I faced the very serious problem: Most media files could not be automatically verified! As said previously Scandisk is useless in this situation unless the whole system experienced crashes. One needs to verify each file for errors specific to the file format with decoders that will report errors instead of concealing them!

The operating system and applications take up insignificant ammount of disk space in a clean, streamlined installation. With some work they can be restored to a working condition. What you have to worry about is your library of video and audio recordings, books, family photosm, or your work for other people.

Now to make my post useful for other users, here is the command that can help a little with video file verification:

ffmpeg -v 5 -i MEDIAFILE -f null - 2> STATISTICS

You need the command line version of ffmpeg. With this command it will decode the passed MEDIAFILE and output STATISTICS to a text file. The statistics will contain sort of decoding progress bar and what’s important most (if not all) errors the decoder has attempted to conceal. The progress information can be easily skipped visually. Each mediafile needs it’s own statistics file.

I can only recommend that you use formats that allow verification: lossless compression for audio (FLAC, Monkey’s), pack your images in RAR (w/o compression), etc. Matroska may also have error detection, but I’m not sure.

I completely agree with your technical description. But I fail to see how this situation can be considered lucky in any way…


#17

You RAM status has no effect on the well bing of the content of your hard drive.


#18

This is not correct.

There are some ways in which bad RAM can damage files that are written or updated, or even corrupt a filesystem as described above.

The larger the percentage of RAM that is bad, and the more you write to the disk while you have bad RAM, the higher the likelihood that some file(s) will eventually be corrupted.