Problems with corrupted images

Sorry for my english its by far not perfect....
which image software are you using now?

 

Brandis,
Can't mention other imaging software in a forum that's owned by Acronis. It's not important anyway; the point I'm trying to make is that the other program makes images that are verified and viable while Acronis produces corrupt images on the same equipment. It was a rebuttal to Tsu's assertion that the corruption problem is caused by something other than Acronis Trueimage itself.

 

I think it's fair to tell us what other software you found that works without corruption so that others can test it on their hardware if they want.

This will help Acronis get to the bottom of the problem.

John

 

I agree, if Acronis doesn't like it then I suppose they can delete the post.

 

Klattu no problem, send me a mail sa@tronicpool.com I promise I will not tell it to Acronis.......

I dont know if you did read the tests that I did on my pc (you can read them in this thread, they are the huge ones with the bad english ), but if you take a look you will see that I could "produce" again and again corrupt files on my system only by copying(copy is a W2000 process and hase nothing to do with Acronis TI) the files from my PC via USB20 to my external MAxtor XT5000 250GB drive. At the start I did this only with True Image files then I stoped and created files using WINZIP (WINZIP hase nothing to do with Acronis TI) .
I dont know why you did have problems with TI and why the unspeakable other image software is working well on your system. Acronis TI can have errors (like every software) but till now it helped me many times.

 

Hello, are you prepaired for some bad english??

here I am with a new test!

Now I did check if the large file corruption appears if I use 1394 Firewire!
My test pc at home:
Athlon64@3200, Asrock Z8 comboboard ALI-chipset, 512MB, ATI9800pro, 120GB P-ATA UDMA5, W2000SP4, USB20 onboard, ultacheap FIREWIRE card from Conceptronic 4 ports in a pci slot.
The external drive is a MAXTOR XT5000 250GB NTFS with USB2.0 and 1394 FIREWIRE.

HARDWARE is the same like the test I perfomed with USB20.

1. I check the MD5 of the file on my ATA disk (file size 3.6GB)
EE7CBE9788B18483239594B539069CBD
2. I copied this file from 120GB ATA via 1394 on the Maxtor
3. I check the MD5 of the file on the MAXTOR
EE7CBE9788B18483239594B539069CBD it is O.K.
4. I copy the file form the MAXTOR via 1394 to the ATA disk (other partition)
5. I check the MD5 of the file on my ATA disk
EE7CBE9788B18483239594B539069CBD it is O.K.
6. I copy this file from 2.ATA partition via 1394 to MAXTOR in an othe folder like in step 2.
7. I check the MD5 of the file on the MAXTOR
EE7CBE9788B18483239594B539069CBD it is O.K.

So I did not get any corrupted file using 1394IEEE-Firewire with the same hardware. On my system 1394IEEE seems to be the one an only choise.

(On my Acer TM632 Labtop I use allways the 1394, because it hase only USB1.1, and I did NEVER hade any corruption problems)

The transfer speed that I did get:
ATA to external drive via 1394: 150Mbit/sek
external drive to ATA via 1394: 245Mbit/sek

Only for comparison again the USB20 transfer speed:
ATA to external drive via USB20: 200Mbit/sek
external drive to ATA via USB20: 164Mbit/sek

Interesting, 1394-Firewire is slow writing to the external disk and significant faster reading from it. With my USB20 onboard I have the opposite.

 

Hi Stefanos,

Another good post!

The key is that you are using everything the same except that the USB 2 is on the motherboard, and the Firewire is on a PCI card. Of course, the cables are different also, but it looks like the USB chipset or it's drivers is the source of the problem.

Is the USB chipset from ALI?

John

 

Klattu,

Thanks,

John

 

Yes the ASROCK Z8 hase a ALI chipset! I will have a look if there are some new USB driver. Also ASROCK hase a new BIOS available, but they want me to make the flash using DOS, and I didnt have time and the guts to do it!

 

OK, I have the same problem. Using the True Image 8 trial. Made an image of my c: drive (disk 1) on the d: drive (disk 2). Image on d: shows as corrupt.

I made an image of c: onto the same drive. This images shows as NOT corrupt.

Copy this image on c: to d:. On d: it shows as corrupt.

I am using the check image feature. These drives are just IDE drives on the same cable. c: is master, d: is slave.

md5sum the original image on c: and the copy on d: and the hashes are different.

The image is about 37 gigs. I will try splitting them up into smaller files. I copied an 87 meg exe (test file) from c: to d: and md5sum reported them as the same.

Charles

 

Hello Charles,

your test file on 37MB is to small. Make a biger testfile. You could use WINZIP and ZIP as many files you find to get at least 1 GB. Make copies of this file from C ---> D and check the MD5 at least 3 times.
Copy the file back from D--->C in a different folder and do the same!

The difficult is to find at what file size the corruption start, with my home PC 300MB are ok. 600MB get corrupted (pc to external disk via USB20)

 

I will try some different size test files tonight with the md5sum program. You have to run it 3 times on each? So between times you run it, the damn test file might change? Yuck.

BTW, I'm running Windows XP SP2, NForce4 motherboard, PerfectDisk 7 for defragging, nothing else I can think of that would mess with files.

If the corruption can be proven to not happen at a smaller size, I'm perfectly willing to split up images. I'd prefer not to, but I'll live with it if I have to. I, too, used another image program (think of the color yellow) that I am fed up with...it had the same problem with corrupted images (and other problems).

 

I see that there are a lot of file corruption problems which may or may not correlate with TI, which may or may not correlate with USB drives, and which may or may not correlate with jumpering drives to Cable Select. Certainly I have seen reports that external USB drives tend to corrupt data.

While everybody is quoting TI's opinion of the file (corrupt or not) and checksums, nobody has reported a test which might shed light on the subject (and might also be a waste of time): why not make binary file comparisons of good and corrupt files? There is free software to do this; I use very old MSDOS programs called CMP.COM and BFC.EXE, but I don't know if they handle large files. The sort of questions to consider: where does the corruption occur? Is it always in the same segments of the files? What is the nature of the corruption: a bit changed here and there, blocks of 256 NUL bytes starting at an offset an exact multiple of 256 from the start of the file, etc. I would not expect the size of the file to change.

For information, I have saved and restored drives a few times, always using a boot CD, from and to a Gigabyte 8KNXP (Intel silicon) motherboard and a Fujitsu-Siemens Amilo laptop using a Freecom 1.8" USB2 drive, and have never had any trouble (yet?). Total sizes sometimes about 2GB, sometimes about 16GB compressed.

HTH

 

I repeated the corruption test last night with a 480 meg file. Copied it to the d: drive. Then hashed the original three times and the copied file. They check out as the same hash.

Regarding the last post, anyone know of a binary comparison tool for huge files (40+ gig)?

 

No the test file dont change, but IF you have reading errors then the result will be different EVERY time.

If you have only writing errors the file will be written wrong, the MD5 will be different from the original, but you will see every time the same wrong MD5.

So repeating the MD5 calculation you can find out what is going wrong!

 

480MB is o.k. ? then go to 1GB.... of cause we all have nothing better to do with our time.....

 

I did the same as your quoted part, with a file that came out to 3.2gig. Same thing happened, if you back up to the drive you are imaging everything is ok, backup to a different drive and it is corrupted.

 

Did another check. Booted using the Cd and made an image of 3.2gig to another computer on my network. Verified and it was ok. Booted to XP and re-verified and it was still ok.

 

KCXLT,

You said that if you backup to a different drive, the image is corrupted. Is this another partition on the same hard drive or a second internal drive in the same computer?

If you copy the "good" image on the same drive that you imaged to the different drive, is the copy still good?

If you boot from the TI CD, and make the image on the different drive (not the network drive as you reported) is that image also good?

John

 

Hello all,

I don't possess an external USB drive but have been following this thread and the one titled <Who DOES NOT have a problem with corrupt images...> with interest. Unfortunately it's becoming more difficult to follow what's going on because many users, but not all, are posting in both threads.

Perhaps one of our Forum Administrator/Moderators could see fit to lock the other thread and direct that the discussions be continued in this one?

Regards

 

hi all,

i found this discussion while searching for a solution for my similar problem.
i would like to share my observations with you on this issue.

i ran Ti 7.638 for about a year on several pcs and laptops and never had problems with copying backups between different disks or via the lan.
also restoring via the lan was working as expected. while os was installed on
ide-hdd, data was stored on sata-hdd.

recently i changed harddiscs, and thats when it all begann. first it was my on fault that w2k sp4 needs reg entry to support big lba. ( xp sp2 entry is default )
so i lost a lot of data above 137gb. after setting this 48-bit reg entry this sudden data death was stopped. but trying to restore for previous backup failed.

creating new backup while running w2k and saving to second hdd ( 160 gb hitachi ) was fine. check was ok. starting form boot disk and checking the backup reported corrupted image. the image was not moved between creation and second check.

i then tried to created new image while running from boot cd. during the backup process i decided to stop the backup-process because i had another idea. the software asked me if i was sure and then stopped the process.
after starting w2k again, i had to realize that the first 60gb partition on my second hdd was gone.....

before i started TI from cd that partition contained 30gb of data!

so i am really frustrated and do not now how to proceed without fearing
to loose more data, especially if you thought to be on the safe side due to several backups....

any ideas? is this may be linked to problems with the TI drivers and big hdd's?

thanks ibuerfei

 

1 question and 2 remarks

Question: Is there a known way to view, and maybe extract, files from a corrupt .tib?

Remark 1: It seems that images that were OK at first can become corrupted after some time. I succeeded in recovering C:\ two times OK, the third time the image was reportedly corrupted, and to my knowledge, nothing had happened to it exept reading it/restoring it.

Remark 2: I feel there is an error in the sequence when trying to restore an image. It seems the partition is first deleted, but if TI subsequently finds the image-to-restore is corrupt, the partition remains deleted. Happened to me twice. (Recoveryexpert helps, but this should not happen: first check if image can be restored, *than* delete original partition).

(But I'd like to *stress* my question....).

VWIW,

Hennie

 

I'll have to do some checking, I should write it all down so I don't forget.
What I did check is the image I made from the CD, saved to a network computer, verified with the CD and within XP was ok. I just copied that file back to my computer, to a different drive than the image was created and verified it, and it was ok.

Does anyone elses progress indicators in verify mode screw up. My current progress bar always stops about 75% of the way across, while the total bar goes all the way.

 

Progress bars do the same thing for me too. I never worry too much about them.

 

Hello, this message is for ALL USERS WHO HAVE CORRUPTION PROBLEMS with internal or external harddisc.

To get more valuable info I would suggest that you download
Programm HD-TACH version 3.0.1.0 www.simplisoftware.com
Make a long bench test (32mb zones) with all your drives (if you have more then one you can compare the results in one screen)
and post the results here in the thread
https://www.wilderssecurity.com/showthread.php?t=65129
making a picture (GIF format) of the results like I did.
Everyone can add picture attachments (look below in "Additional Options" press "Manage Attachments")

Perhaps we can isolate the problem this way!