HDD Data Corruption with RAID !PLEASE HELP!

[D75]

PLEASE HELP!
I have a problem with my newly purchased RAID card and harddives.
Iv seen 2 posts on the board that maybe addresses the same problem, but no solution is given.

This is what i first purchased:
- PROMISE SX4000 RAID Controller. Handles RAID0, RAID1, RAID0+1 And RAID5
- 3 x IBM/Hitachi 120GB 7200RPM IC35LI120AVV207

After flashing a new BIOS to the RAID controller, I installed the controller and drivers in my windows XP.
I configured a RAID 5 with my 3 harddrives.
I copied everything from my old 100GB HHD to this new RAID drive after formatting it with NTFS.
After a while I noticed that some of my compressed and other large files were corrupt, and lucky as I was i discoverd this BEFORE erasing my old HHD.

Now, after realizing I have a problem, I tried the following:
- I started comparing the corrupted files with the files on my old HHD, and found that about 20 bytes in 500MB were corrupt.
- I then reformatted and tried to copy the data again, but it was always the same problem.
- I tried all possible RAID configurations, but it was still the same problem.
- I tried my onboard RAID controller, that handles JBOD, RAID0, RAID1, RAID0+1, but with the same result.
- I then purchased a NEW "standard", "non-RAID" controller, but it STILL was the same problem.
- I then tried DFT using this new controller and it said everthing was fine with my 3 harddrives.
- I tried my 2 controllers and 3 harddrives on another PC with the same results.
- Finally I tried my integrated onboard standard non-raid controller, and it works fine!!!

Ofcourse I want to make my RAID controller work as It should, but I just cant find the solution!

Anybody have any ideas?
I dont think the problem is ALL my 3 brand new harddrives, that DFT says works fine.
Seems strange that the problem should be my 3 controllers, the purchased RAID, the integrated RAID AND my purchased non-RAID controller...

I dont know what to try next!
PLEASE HELP!


<P ID="edit"><FONT SIZE=-1><EM>Edited by D75 on 05/16/03 05:29 PM.</EM></FONT></P>

 

[lhgpoobaa]

hmmm. all 3 drives DFT fine...
I assume you are using the latest drivers for the raid card yes?

It could be that one of your cables is slightly damaged, or is recieving interference, or has a connection issue.
They way i test things like this is to make a big compressed archive file then test extract it, looking for data corruption errors. So possibly do that sequentially with each cable and try to find the bad one. (if it exists)

<b>Melb_angel = THGC's <i>INNOCENT</i> Angel</b>
<A HREF="http://www.picturetrail.com/master_poobaa" target="_new">PooBaa's Pics!</A>

 

[D75]

It cant be the cables, Iv tried 9 cables. 3 different brands, 2 different types.

The way of testing you describe is exactly the way I do it.

Iv got another thing im going to try. I think that the problem occurs when I copy files between harddrives on 2 different controllers. It occurrend to me that the only way I can get it to work is when a place a drive on the integrated standard controller, and that is the same controller that I have the drive with my old data, the one I copy from. Im going to try to put my old drive and one of my new ones on another controller, and then copy.

I have not tried this yet, and Im running out of to try, so it seems a bit far fetched.

Ill post my results later.


/D75
"Doh, Though I knew all 'bout hardware"

 

[D75]

Iv tested it now and its the same problem. The only combination that works is when I place the drives on the integrated controller.

I just copied a whole 100GB partition with the integrated controller using partition magic 8, and compared the whole volume. Not a single byte that didnt match.

I just got another idea. Im goin to try another harddrive, and place it on the PCI IDE controller, and see if it works. Its a completely different model.

Im beginning to think its a chipset problem or something.


/D75
"Doh, Though I knew all 'bout hardware"

 

[unoc]

I have the same controller driving 4 IBM 180GXP with RAID 5. I never had any problem even modifying PCI frequency.
I copied files as big as about 600 MB each (disk image).
Have you installed the last BIOS/drivers ? 1.0.15 and 1.0.16.
What kind of mobo do you have ?
Have you enabled the PCI bus master ?
Have you installed the last drivers for your mobo ?
However, if you have problems even without using the SX4000, your bag is definitively to be found into the mobo driver, very probably into the PCI bus driver.


______________________
<font color=red>the new bios of my mobo let me to choose the P.rating number of my CPU.
Now I have an XP 8000 + </font color=red>

 

[D75]

Yes, Iv got the latest BIOS and drivers.
My motherboard is a GigaByte GA-8SQ800 Ultra. I dont know the model of the other modo iv tried it on, but im going to check that later and post it here.
Yes, Iv installed the latest mobo drivers.
Enabled the PCI bus master? What is that? Where can I do that?

I havent had time to try the other harddrive yet, I hope I can do that later today. Im begining to suspect its a driver problem myself, or maybe a chipset problem, but I dont know how to fix it. Iv e-mailed GigaByte, but you never know when they are going reply.

Iv also got an offer from the company where I work to exchange the 3 IBM drives with 3 WD to test. But I want to be sure its the drives first. But I dont think it is. Iv read in this forum that WD is better, so maybe Im going to do it anyway. What do you think?

Anybody else out there have the same mobo with a PCI controller card?

/D75
"Doh, Though I knew all 'bout hardware"<P ID="edit"><FONT SIZE=-1><EM>Edited by D75 on 05/16/03 04:44 AM.</EM></FONT></P>

 

[D75]

OK, now iv tested it with 3 types of harddrives, and its still the same problem. It seems I get this error when I use a controller on the PCI bus. Its trange that even the integrated RAID controller has this problem!! I think that actually is connected to the PCI bus too.

Gigabyte hasn't replied to me e-mails yet, and you never know if they ever will...
What the ?!#¤#@! am I gonna do!?!

Damn, im getting tired of this...

PLEASE HELP ME SOMEONE! =P


/D75
"Doh, Though I knew all 'bout hardware"

 

[unoc]

When you install a PCI RAID card you should disable the on-board raid. Have you done so ?
I have the same controller. When I installed it I had no problems. Try to e-mail Promise. They are very kind and particularly for the SX4000, a new product, they are really speedy with support.
You will have the PCI bus master enabled after the installation of the SIS chipset drivers, <A HREF="http://tw.giga-byte.com/Motherboard/Support/Driver/Driver_GA-8SQ800.htm" target="_new">http://tw.giga-byte.com/Motherboard/Support/Driver/Driver_GA-8SQ800.htm</A>
The bus master allow you a faster PCI bus without charging the CPU of extra work to manage the PCI data transfer.
It needs also you install the Service Pack 1 for Win XP that allows you to recognize better your hardware.

______________________
<font color=red>the new bios of my mobo let me to choose the P.rating number of my CPU.
Now I have an XP 8000 + </font color=red>

 

[D75]

I always disable the on-board RAID when testing the SX4000 raid card.
Iv got 2 Computers with Windows XP. They completly different motherboards. I get the same problem, with ANY PCI IDE controller.

I install the mobo drivers, but it make no difference. I have all the latest updates to both Windows XP installation. I have even tried a completely new installation of Windows XP.

Iv sent Promise an e-mail, but I got no answer.

I tried an ME DOS boot disk, copied a large file and then did an FC to compare, and then it was no problem at all!

What OS are you running? Im beginning to suspect its a windows XP problem. A friend of mine had had the same problem with a RAID card and XP, so he runs Windows Server 2003 on a separate file server instead. Im going to try Server 2003 and see if the problem remains.




/D75
"Doh, Though I knew all 'bout hardware"

 

[D75]

I tried Windows Server 2003 with the same problem.. Damn.
Right now I cant come up with anything more to try...
And I still havnt got any reply from any support at Promise or Gigabyte.


/D75
"Doh, Though I knew all 'bout hardware"

 

[unoc]

I installed a dual boot Win2K SP3 Workst./WinXP SP1 Professional (English version)
The HDD array is composed by four IBM 180GXP 60 GB which constitute a RAID 5 array (3+1 180GB+60GB for parity).
The Sx4000 BIOS is ver. 1.01.0.15
The WinXP driver is ver. 1.01.0.16
The remaining component are:
PSU Enermax 465 W
CPU AMD Athlon XP 2700+
Mobo A7N8x Deluxe PCB 1.04 bios 1004
Promise SX4000 Controller +256 MB SDRAM PC133 on PCI slot 3
AGP card Sapphire Radeon 9500 128 MB
ISDN modem on slot 5 (the last one)
ONboard sound
USB 2.0 enabled
DDR RAM Corsair extreme 2x512 MB (dual channel enabled) CAS 2 PC3200
IDE1 IBM 120GXP 120GB master IBM 120GXP 120GB slave
IDE2 Yamaha F1E burner master ASUS 52x CD drive slave
I tested the system with FSB up to 200 MHz without any problem, the CPU works fine (air cooled) up to 2430 MHz/49 °C
The RAID controller works rock solid with both OS Win2k and WinXP.
What slot did you use for the Promise card ? Have you tried to change the PCI slot ?
Have you tried to change the system memory and the controller memory ? (SDRAM you have to put into the controller card)


______________________
<P ID="edit"><FONT SIZE=-1><EM>Edited by unoc on 05/20/03 08:27 AM.</EM></FONT></P>

 

[D75]

I think everthing checks out ok.. On both my computers Iv tried it on.

Iv tried the Promise RAID on slot 1 and 2, and the normal controller on 1,2 and 5.
The integrated RAID.. well, I guess I cant move that one =)

I have not tried to change the system memory since I havnt got any other memory. But I doubt thats the problem, because it works with the onboard normal controller, and Iv got the same problem with 2 different mobos with different memory.

I have tried 3 different memory modules for the Promise card. 2 64MB and 1 128mb. No difference.

In the BIOS iv got "default" settings with no overclocking...

Since I have the same problem with the integrated RAID, I hope GigaByte will answer soon and maybe have a solution. I dont think its a problem with ANY controller, its got to be something else, I just cant understand what it is.




/D75
"Doh, Though I knew all 'bout hardware"

 

[unoc]

Are the boards you tried the same model from Gigabyte ?
If yes it could be a bug in this particular mobo bios/hw.
Because of these repeated failures, can you substitute the board with one from an other brand ?

______________________

 

[D75]

The other mobo is an Aopen AX3SP. I think it has VIA chipsets, while mine has SiS.
Im not 100% sure bout the chipset on the Aopen mobo though.

I cant figure out if these have something in common, can you?



/D75
"Doh, Though I knew all 'bout hardware"<P ID="edit"><FONT SIZE=-1><EM>Edited by D75 on 05/20/03 03:24 PM.</EM></FONT></P>

 

[unoc]

The AX3SP mobo should be an Intel 815 based mobo for P3 and Celeron.
Hmmmmm..this is quite an old mobo and still does not assure us that the problems do not come from the chipset. However it is really difficult that the problems come from the mobo.
Summarising:
Motherboard..................OK
System and controller RAM....OK
Hdds.........................OK
Slot choice .................OK
OS (WinXP or Win2k)..........OK
Controller...................OK (I have the same and it works fine)
Files........................OK (they seem to be not corrupted at the origin)
PSU..........................OK (It generally influences the CPU)
What can you check more?
Is your computer grounded ?
The ground should be delivered with the power cable.
Touch slightly the case of your computer with the back of your hand with your feet on the floor. If you feel a little tingle, the computer is not well grounded and electric noise may corrupt data transfer along the cable.
I cannot suggest you more tests but to buy a new motherboard.

______________________

 

[D75]

Yes, I belive its grounded. That both my computers are.

Now, listen to this:
I tried it on 2 computers at the company where I work. HP Vectra P3.
They were a slightly differnet model, and one had Windows 2000 and the other Windows XP.
My test worked on both of them!!

So, this must mean its a mobo problem.
Now, I wonder. Does AOpen and GigiByte mobos SUCK BIGTIME (excuse me, but im getting a bit irritaded =P ), or is it so that I have so EXTREME bad luck in my mobo purchases?

The thing is Iv got another AOpen mobo, An AX34 with a P3, and I think Im going to try it on this. The reason for me not to try this mobo before is ofcourse I though 2 mobo tests have got to be enough... Well, If it works, I guess ill put together a fileserver! =P

Now iv got a couple of hours work to put the mobo together an all.
Ill report back here later.



/D75
"Doh, Though I knew all 'bout hardware"

 

[D75]

OK, I tried to test it with my old AOpen mobo, but apparently its broken.. talk about bad luck..

Anyhow, I got a HP Vectra VL P2 from my workplace to try it with, and it worked great. Im placing that computer as a fileserver.

The only problem now is me trusting that RAID array...
That was the reason I bought it in the first place.

Trusting AOpen or GigaByte ill never do again! =(

Thanks for all your help!




/D75
"Doh, Though I knew all 'bout hardware"