HDD Failure makes no sense

Hi,

I was performing a routine backup when I discovered an anomoly which appeared to be a faulty hard drive. The backup was to a pair of stripped 120G western digital drives connected to a rocket raid 1620 controller, which appears to Windows as a single 240G drive P:.

Oddly , after the anomoly, I was able to access certain directories on P: without problems but others would cause explorer to report that the drive was 'not connected'. After this message was displayed, access to any directory on P: would report the 'not connected' message. A reboot however allowed access again to certain directories until the 'trigger' directory was accessed and so on.

I have done a lot of investigation into the problem including changing cables, swapping RAID channels etc. but the fault always follows the pair of drives.

So next I connected each of the 120G drives in turn, to a spare motherboard IDE channel and performed a normal (long) format followed by a full diagnostic test in order to locate the faulty drive. So far I have managed to load about 100GB of data to one of the drives but ran out of time to do the same wih the other.

I was surprised to find that for both drives there were no problems?

Next I reconnected the two drives back into the RAID controller, created a new RAID strip, a new Windows partition/logical drive, IE P:, and then performed a normal format. This time Windows reported that the format did not complete successfully. I then created a new strip using the other two channels of the 1620 RAID controller to elliminate the controller itself. This again reported the formatting problem.

So now I am wondering how two drives can format and test OK separately but fail when stripped. What is the best test I can do to elliminate a fault with the hard drives?

Any help or suggestions would be greatly appreciated.

Regards Basher.

 

Let the fun begin ..........

You may want to try going back with the drives in their original position re: boot and hidden and then try deleting the RAID Array first, then FDISKING MBR and then FDISKING each drive and then formatting each drive but that might not kill everything on those drives. I've found it necessary to "clean" drives like this in a totally different machine, physically jumpered as slaves and don't pretend to completely understand why this is necessary. Just know what has and hasn't worked for me.

 

I will give that a try, thanks.

After I sent the post late last night I come across information regarding the differences between a standard 40 pin EIDE and 80 wire 40 pin cable that supports the faster ATA100/133 drives.

I may have caused this problem my self which I am looking into now. You see I have 80 wire 40 pin cables plugged into the hard drives but the cables were not long enough so I made some standard 40 pin ribbon cable extenders.

Could this cause a problem?

Regards Basher

 

Not only could, but would ..........

 

Yep it was all my doing.

I did not appreciate the differences between the 40-wire and 80-wire 40-pin connectors and of course the 80-wire section was telling the motherboard that it could operate at high speed but the extender itself was not up to it!

All sorted now. Strange, though, how it worked for quite some time as it was until recently. Oh well ya live 'n' learn.

Regards Basher