Having a REALLY bizarre XP Pro install issue SOLVED

[Ardan]

I just thought that I would ask here first, since i'm unsure. I'm going to be very thorough since it seems so strange to me. I'm betting it is a hardware issue, but I don't know for sure. Anyways, here goes:

I have a setup that has been going rock solid since I got it (bought the 90nm winchester Athlon 64 3000+ when it came out) and it serves my needs.

AMD Athlon 64 3000+ @ 2.4Ghz on an Abit AV8
1GB (512x2) Crucial Ballistix DDR400
2x80GB Hitachi Deskstar 7K250 SATA HDDs in a RAID 0 array (no, its not for 'games')
GeForce 6600GT (overclocked too)
420W OCZ PowerStream PSU
XP Pro 32-bit on the RAID array, Gentoo Linux on a 40gb PATA drive

Anyways, the problem is not with my PC. A co-worker is getting into the same stuff (he hasn't started with video or audio or anything else, so what he wants to do IS irrelevant) and he was going to build a setup identical to mine since he really likes its performance and stability. Has NEVER failed me (I do backups to a 250GB external drive though) either.

He built and showed me the computer...looked like he cloned mine instead . He bought the 80gig drives from someone that had been previously using them. He grabbed a trial version of XP Pro 64bit to try out first, and so he didn't set them up in a RAID array yet. He noticed that when he went to install, the installer was telling him it couldn't copy some files over. He said they were fairly random which ones when he'd try again and again on failsafe settings, and it would freeze up instantly once it finished the install and went into XP Pro 64bit.

I came over and checked it out and one message he neglected to mention was that it was warning him about a disk driver wasn't found or properly installed. I do not have the exact message, but it happens each and every time so I could get it for you if you wish.

I took his store-bought copy of XP Pro 32-bit and wiped both drives clean (just in case) and it was also failing to copy over some files. Let me clarify that it was before the actual GUI install that it was showing those messages. You could either retry, skip the file or quit. One was an exe once, another in the same install was a .lex file, and they seemed to be quite random. It also showed the same disk driver message once, but none of the other times. Windows was completely unstable each time, however.

I took his XP Pro 64bit eval and put it on the PATA drive in my PC and I had no issues installing it. I figured it was the HDDs, but I put my 40gig drive in his PC and attempted both installs and I got the same messages as his drives. Anyone have any ideas? If you want to know exactly what files they are, I can provide you their names. I'm just quite mystified at this point and thought I better ask first. He has a new laptop, so its not an urgent problem. He doesn't rely on the PC for everyday tasks.

I have ran Fedora Core 4 on my 40gig drive in his PC successfully, have not tried in the drives he bought. The motherboard shipped with the latest BIOS version, which I find to be very good. The SATA/RAID drivers he put on a floppy came off of Abit's site and were version 4.20 (latest is 4.30H I believe).

If this is a hardware problem, please don't blast me for it. It is now 1:27am, I've worked on it for 6 hours and we dealt with software all night . Well, I tried my best to be detailed...hope it helps.

 

[tyanni]

Have you tested the ram? Random errors when installing windows, especially if the hdd's and cd have tested fine, are often caused by bad memory. I'd try running memtest on each stick individually and testing each slot for a while (at least a couple hours if not overnight) to rule this out. Also, it could be the actual cd drive in his machine, have you tried putting that in yours and installing windows from it?

 

[spherrod]

Originally posted by: tyanni
Have you tested the ram? Random errors when installing windows, especially if the hdd's and cd have tested fine, are often caused by bad memory. I'd try running memtest on each stick individually and testing each slot for a while (at least a couple hours if not overnight) to rule this out. Also, it could be the actual cd drive in his machine, have you tried putting that in yours and installing windows from it?

Yeah - that kind of issue sounds like memory, as tyanniy suggested run Memtest overnight and see what results you get

 

[bsobel]

My usual suspects:

Ram
IDE cables

Bill

 

[Ardan]

Good idea, I didn't think about testing the RAM. It makes perfect sense...but it was almost 2am when I wrote that

The drive, an ND-3500, is pretty solid but it is a good suggestion and we'll test that out. Throughout the GUI install attempts, it would miss a file and then when you browsed to where it thought the directory was, the file was listed there but it would never copy the file over.

i'll swap out the cables and test the ram and hope the cables are bad and the ram is okay .
I'm sure he would be bummed to buy some Ballistix ram and find it is bad right off the bat, but its possible.

 

[Ardan]

An update. I took out his brand new NEC ND-3540A and put in my ND-3500A and tried to install and it worked great. I put his and a HD of his in my PC and tried to install, but it didn't work. I got through and when it was loading files to get only to the HD partition section before the GUI part of the setup, the LED would just stay lit and the drive would just make a clicking noise about once every 3 seconds. It would then, after maybe 30 seconds of doing this, give a BSOD. I noticed it wouldn't load select files when it was clicking like that. I would skip the file and it wound up the CD and would transfer more before it did that again.

I did the mem test and the advanced disk tests on both SATA drives with hitachi's DFT utility and checked out the cables and it all turned out okay. Everything appeared to be in top shape until we swapped out the ND-3540A. It really bummed him out that it came a little screwed up and he is obviously going to be contacting NEC to get a replacement. He chose it after reading about all the praise of those ND-3500 and successive burners from NEC on Anandtech too, you know. I like mine a lot, and its too bad that one was messed up

 

[imported_Phil]

Originally posted by: Ardan
I got through and when it was loading files to get only to the HD partition section before the GUI part of the setup, the LED would just stay lit and the drive would just make a clicking noise about once every 3 seconds. It would then, after maybe 30 seconds of doing this, give a BSOD. I noticed it wouldn't load select files when it was clicking like that. I would skip the file and it wound up the CD and would transfer more before it did that again.

I hope that those hard disks will be replaced, as that clicking noise is the first signs of imminent failure.

 

[Ardan]

No it was the DVD drive that did an unusual clicking noise, not the hard drives. We called NEC for the drive and apparently the noise is tied into its malfunction. They wanted us to send it back for a replacement or fix for it (it can be fixed I guess) but Ideclined. This doesn't put a damper on my high regard for their drives, though. His case sounded quite unique.

Curiously, we got the Benq DW1640 replacement burner from newegg and he still had troubles after we thought everything was great. He is now ditching the AV8 for an EPoX EP-9NDA+ or whatever model number it was. Its an nForce3 Ultra board and it was reviewed on Anandtech once, so i'm sure 99% of you know which one already . I had him take the box to a reputable repair shop in MN for a 2nd opinion and they said the NEC drive isn't working properly and the AV8 has a serious issue with it.

Now, I never looked at it but they showed us a shot of his IRQ info. It went like they are usually, 0...1...2..etc. However, it skips a few #'s and goes to 64, 65, 68 and the 6600GT is at an irq of 128. Every peripheral worked fine for them (except that NEC drive) in other systems, just not his. The VIA Velocity gigabit ethernet couldn't acquire a DHCP signal or do just bout anything and its IRQ was '64'...the SATA drives were losing data (RAID or not) and their IRQ was 68. Also, heavy graphics use was met with a frozen screen that wasn't completely frozen (he could alt-tab out of the frozen dxdiag 3D test) and the IRQ for the video card was 128. Also, it said the motherboard wasn't supplying enough information for either a linux or a windows system to make use of the firewire connectors, and that IRQ was also 64. They said they also cleared the CMOS and flashed to the previous BIOS version, 23, but it wouldn't flash completely. I witnessed it, and I must say that this board's problems are baffling.

I am not sure if the two failures are coincidence or tied together, but I do know that its a good enough reason for me to do a switch. I bought the same motherboard online today myself with next day shipping. With all the BIOS issues i've had (but no real instability), his mammoth problem is enough for me to switch. I really don't think mine will have a problem like that, but since it is no inconvenience I think I will anyways. I have a good enough reason I guess: every time I flash the BIOS for an update that fixes a problem I had, another would pop up! Like they fixed the S3 resume problem and I magically couldn't see my SATA drives because USB Emulation was on. Eh? weird.

Hope your week was better than his!