Using USB to IDE/SATA adapter gives different MD5SUMS.

[ironk]

Using a USB to IDE/SATA adapter, and it causes some strange behaviour. I have the adapter connected as follows:

Desktop computer > USB Adapter > IDE Hard Drive.

I copied a file from the desktop computer to the "usb" IDE hard drive, then ran a MD5SUM. The checksums came out different.

Then, if I connect the same IDE drive directly to the desktop computer as slave and run md5sum, the checksums are the same this time.

Anyone else run into this problem? Are checksums supposed to be different when using a usb adapter? Why do they verify correctly if I connect the drive directly?

The adapter I am using:

http://www.newegg.com/Product/...tec%20usb%20ide%20sata

I have tried the adapter on different Windows computers, same thing happens. Even used different MD5SUM software, came out with the same result.

Thanks.

 

[ironk]

bump

 

[ViRGE]

The MD5SUMs should absolutely not be different. A different MD5SUM means the file is different, which means it's corrupted. Trash that adapter, stat.

 

[ironk]

But why do the MD5SUMs check out fine if the hard drive is connected as slave and not as a usb?

What brand/model adapter do you guys use? Any recommendations?

 

[ViRGE]

It's probably getting corrupted when being read via the USB adapter.

 

[ironk]

That means that the usb adapter is fine, but there something wrong when MD5SUM is checked via a adapter. That is really strange.

I was thinking others would've experienced this by now, going to look at the newegg reviews a bit more.

 

[0roo0roo]

well its not fine. no difference should exist.

 

[ironk]

anyone else ever check the md5sum of file while its connected to an adapter? I wonder if the jumper settings on the HD have anything to do with, I have it set as cable select.

 

[Aluvus]

Originally posted by: ironk
That means that the usb adapter is fine, but there something wrong when MD5SUM is checked via a adapter. That is really strange.

It most likely means that when the hard drive sends data to the computer through the adapter, bad things are happening to that data. That does not mean that the USB adapter is fine; it in fact means the opposite of that. The USB adapter is somehow mangling data that is sent through it. It should be passing on the data so that when the CPU looks at it to calculate an MD5 sum (after all the extra protocol stuff has been unwrapped), it is identical to the original data.

If you opted to write data to the drive through the adapter and then read it, you might get even more interesting surprises.

You have an adapter problem, not an MD5 problem. Toss the adapter.

 

[0roo0roo]

yup or keep it.
quiet corruption is fun if you like living dangerously the thrill!

 

[ironk]

If you opted to write data to the drive through the adapter and then read it, you might get even more interesting surprises.

I don't think people are understanding as to what I tried to explain, probably my fault. I copied a file to the usb adapter connected HD, and when the copy process was finished, I ran MD5SUM while the adapter with the HD were still connected. The MD5SUMs fail to match at this point. However, if I were to remove the HD from the adapter and connect it as a slave directly to the computer, and then try to verify the same file, the MD5SUMs match perfectly. This is on the same file that was copied using the usb adapter, I just changed the interface.

 

[Mark R]

It indicates the adaptor is corrupting data when read.

The way to test this for sure, is to connect the drive as slave, save a file onto it, and verify it's MD5sum. The connect via adaptor and test it again. If it doesn't match, then the adaptor is causing data corruption, and should be replaced.

Just because the data is getting corrupted across USB, doesn't necessarily mean that the adaptor is definitely faulty. A lot of motherboards from bottom feeder manufacturers have had faulty USB - the major culprit that I've personally come across, has been Abit. I've had quite a few boards from them that chew up data over USB. Every one did eventually get fixed with BIOS updates - but it some cases, it was an 18 month wait for a new BIOS.

 

[ironk]

Yeah, I always suspect a USB/BIOS problem first, have seen usb problems too many times.

But in this case, I tried different computers with different mobos. Same thing.

Anyone try to do this type of transfer/md5sum check recently? What are your results?