Data corruption during network transfer

[Starbuck84]

I have a problem which is bugging me for some time now and I hope you guys could get me some insight in what might be going on.

I have a modem/router from my ISP with a built-in gigabit switch. My laptop, media player and NAS are all connected by wire to this built-in switch. So it's a fairly basic network. My laptop runs Windows 8 x64 and I have a Synology DS212+ NAS (running the latest DSM 4.2).

The other day I found a folder on my digital camera with some images from our holiday. I copied the entire folder to my NAS, so I could show those images on my television using my media player. Everything copied fine according to Windows, except when I looked in the folder, several images were corrupted. Big blue, green or red stripes through the entire image. In the end 6 out of the 168 images were damaged. I recopied those files to my NAS and everything went fine this time... but it kept bugging me. Why were those files damaged to begin with?

Later that evening I deleted the entire photo folder on my NAS, recopied the entire photo folder back from my laptop to my NAS using TeraCopy. TeraCopy showed me several files were damaged. "CRC does not match". Indeed, this time 3 files were damaged and showed the same blue, green or red stripes through the entire image, making those images look horrid. This can be reproduced, every time I re-copy the photo folder, some files (numbers vary) are being corrupted. I also notice the same errors while using SyncBack Free to back-up my data to my NAS. Not every time, but SyncBack does regularly report errors and if I re-run the same profile, everything is fine. And if I use TeraCopy I can re-copy the corrupted files, so in the end all is good, but I lost quite a bit of trust using my NAS for storage purposes. And I'd like to know what might be causing this.

What did I try to resolve the issue?

* Updated my NIC drivers. Using Windows Update drivers and the latest official (WHQL) Atheros drivers. (All settings left standard).

* S.M.A.R.T. (quick and long) test on my HDD in my NAS. (Result: Normal. ERROR=0)

* Memtester (Synology assistant) on my NAS. (Result: ERROR=0)

* Reset my NAS and modem/router.

* Swapped the UTP cable between my modem/router and my laptop. (Direct connection, no wall sockets or anything).

* Tried a different UTP port on my modem/router. (Port 2 instead of port 1).

If I play a movie of let's say 2 hours from my NAS to my media player, everything works flawlessly. Although my media player only has a 100mbit port. Could this be of any meaning?

I was thinking of performing the following tests:

* Copy the same entire photo folder from my MacBook Air to my NAS, unfortunately this has be over wireless. So no real comparison in terms of bandwidth/speed and use a pre-created MD5 to check whether the copy went OK or not. Just to exclude devices.

* Set my NIC on my Windows laptop to a forced 100mbit connection instead of 1000mbit (auto handshake).

* Any ideas are more than welcome!

On a side note, everything I download from the internet, whether I'm using YouTube or download large files works OK. Never had any trouble whatsoever. Just local copies to my NAS seem to go wrong. I never had this issue when I used my laptop wirelessly. I went to a wired solution because of the transfer speed.

Any ideas/thought on the matter?

[The_Decryptor Veteran]

It's quite possible the NIC in your laptop is dying, try over wireless if possible, and maybe borrow an ethernet adapter from somebody to see if there's any errors.

[Starbuck84]

I was afraid of that. If I try wireless, it's all good. The problems started occuring when I switched to a wired connection. I didn't notice it at first, but later on SyncBack started throwing these error messages, and now my photos get corrupted. I could borrow another laptop from a friend. It only has a 100mbit port.

Question. Does a 1000mbps connection offer more stress on your cables/hardware than a 100mbps connection? I'd think so? Reason I ask, is because this is the only machine which runs at 1000mbps. Testing with another machine which "only" has a 100mbps port would be much efficient? Just thinking out loud here.

[The_Decryptor Veteran]

Transferring at 1Gbps will place more stress on the hardware than 100Mbps, so reducing it could certainly help.

You could also try running wireshark and performing file transfers, see if you're getting a large amount of errors show up (Of course your computer could be "smart" and offloading certain parts of data transfer to the NIC, which will show up as errors even if it works fine in the end. Wireshark on my PC thinks every single sent packet is corrupt because the checksum in the header is set by the hardware, so software sees it as "0x0000")

[Starbuck84]

Good tip on Wireshark, thanks! I'll try to borrow my friend's laptop later this evening. I'll also transfer the same photos to my NAS using my MacBook. (Wireless).

I'll create a .MD5 and copy everything to my NAS and check the hash later. If that works, I can fairly certain say the problem is to be sought within my Windows laptop. I'll keep you informed.

[Starbuck84]

After some testing I can fairly say my NIC is the problem in this matter. I've done the following test:

* Create a .MD5Sum of my photo folder.

* Copy the folder to my MacBook and friend's HP laptop.

* Duplicate the folder two times on both machines. (So you end up with 3 different folders containing the exact same data).

* Copy all folders to a temp folder on my NAS.

* Verify each copy using the .MD5Sum.

All 6 folders (in total) verified perfectly. The MacBook used the wireless connection while the HP from my friend used the same wired connection as my own laptop. (Unplugged my laptop and put the cable in his laptop).

Now I read on different forums that it could be related to bad RAM or HDD errors. I'm willing to test those parts without a doubt, but if it were my RAM or HDD, shouldn't I get more problems like an unstable machine while using Windows in general? My thoughts are telling me those parts are OK, could never hurt to test them regardless.

/Update: I even forced my NIC to operate at 100mbps/full duplex instead of 1000mbps. To no avail, the amount of CRC-errors only grew to 26 corrupted files. Guess it's back to wireless for me. :)

Thanks The_Decryptor so far!

[mwc]

How about connecting your laptop directly to the NAS?

BTW, the data corruption first started when coping from digital camera connected to laptop to the NAS right?

[Starbuck84]

I haven't tried connecting directly to the NAS, as another laptop (which uses the exact same cable/connection) is able to copy the same files without CRC-erros. In threefold, see my previous post, while my own laptop can't even complete a single copy without errors.

I noticed the data corruption after I connected my camera to my laptop and copied the data directly to my NAS. But it doesn't matter how I copy the data, later on I copied the data to my laptop's hard drive. Leaving the camera out of the aquision, didn't matter.

/Update: unless I'm overlooking things?

[mwc]

I haven't tried connecting directly to the NAS, as another laptop (which uses the exact same cable/connection) is able to copy the same files without CRC-erros. In threefold, see my previous post, while my own laptop can't even complete a single copy without errors.

Try with YOUR laptop with that same cable/connection directly to the NAS.

This rules out (strange) possibilities

[Starbuck84]

I will try, no problem. Already gave my NAS a static IP, so no worries there. Will report back!

[Starbuck84]

/Update: came home earlier from work. Connected the NAS directly to my own laptop. Had to manually assign an IP address and all was good. I started copying my photo folder, and with the first 100 photos I got 5 CRC-errors. I swapped the cable between my laptop and NAS for a shorter one, placing my laptop almost next to my NAS and I still get CRC-errors. Only one this time, but still, one to many in my book.

[+BudMan MVC]

"as another laptop (which uses the exact same cable/connection) is able to copy the same files without CRC-erros"

Then its clearly your laptop.. Are these pictures on the camera your connecting to the laptop or usb/cf/etc/ stick? Its possible its not your nic in the laptop but reading the files either from your hdd or camera, etc. could be the problem.

[Starbuck84]

You're right +BudMan. The camera is connected through USB, and the camera is seen as a "mass storage device". SD-card. I've copied the photos straight from the camera to the NAS, and I've copied the photos to my laptop's HDD first and then copied them to the NAS. Both copies failed with CRC-errors. So to exclude my RAM and HDD I'll test both of them with UBCD and hope it's just my NIC.

[Starbuck84]

/Update: I've read out S.M.A.R.T. on both HDD's and it reported nothing. Twice I ran a complete surface test on both HDD's using ViVARD (UBCD 5.2.4) and it found nothing. I've let Memtest 4.2.0 ran all night, and it also found nothing. It's fair to say my RAM or HDD aren't at fault here. The only part I can think of is NIC related. I have absolutely no problems or CRC-errors while using Windows in general. Only while copying data to my NAS. Also with all the details regarding my previous posts, I'm unable to conclude anthing else as a faulty NIC. Unless I'm overlooking something?