Need help narrowing the problem down: hard drive failure 2

[psemianonymous]

Let me first start out with a disclaimer: Yes, I do mess around far too much with my home computer and have probably done something I shouldn't have.

Let me tell you of my hardware:
-P3 600MHz with 256MB of generic RAM
-20GB Hard drive (Controller 0: Master) - Quantum fireball
-80GB Hard drive (Controller 0: Slave) - Western Digital
-One DVD-ROM (Controller 1: Master)
-One CD-RW (Controller 1: Slave)


Let me now begin with the symptoms:

1. I first noticed these problems after running Partition Magic and moving/resizing one of my partitions. The process took some 8 or more hours to complete. It (apparently) completed successfully.
2. Red Hat 9 has failed install about 15 times now. I will go into more detail later.
3. Mandrake 9.0 has also failed install 5 or more times. Again, more detail later.
4. Using an old Slackware CD, attempting to copy the contents of the CD to the second hard drive using the simple "cp" command fails.
5. Attempting to copy files from my FAT32 partition on my second hard drive to my first ... fails. On certain files.


Now, you may stop here and say "that's a hard drive #2 problem", and you may be right. BUT.

6. Red Hat 9's "check CD integrity" ... failed on a CD. The CD was verified on another computer on the same day and succeeded. This check was run on both the DVD-ROM and the CD-RW, so it can't possibly be a read error. I have two copies of Red Hat 9 CD's, and both sets fail (but have worked elsewhere).

7. I have a linux swap partition on drive #2.
8. Red Hat's installer tells me my "hard drive geometry may be invalid" or something. I can't remember the error message at this time, but it involved my hard drive geometry.
9. Scandisk (Win98SE) has now failed to "thoroughly" search BOTH my C: and my D:. I have now deleted all other partitions and resized these two partitions so that FAT32 covers all the land. What I mean is that C: is 20GB, my entire drive, and D: is 80GB, my entire drive #2. Scandisk halted on BOTH of these.
10. Specifically what happens is that the computer freezes. During a linux install, Red Hat will poop out after "not finding a package" or after "failing to format a partition". Mandrake install simply halts. "cp" from CD to hard disk #2 halts the computer--unresponsive. Windows explorer "copy" halts the computer--unresponsive. Scandisk fails by halting the computer--nothing appears in scandisk.log


I don't want to go into too much detail about how much I've been pulling hairs. Just please help me with the following questions:

1. Is this just a hard drive #2 problem? Or something more?
2. Could it be my IDE controller, and if so, how can I verify this?
3. Could it be something to do with the fact that I used Partition Magic to mess with my D drive? (remember the red hat error message) I have used all kinds of fdisk: DOS FDISK, linux fdisk, linux cfdisk, Partition Magic (in many ways and shapes), Red Hat's Disk Druid, and hex-edited the MBR to swap two partitions' location in the table.
4. If it is a hard drive geometry/misreporting error, how could I fix this? Is there a utility I can use to low-level format the drive (at this point nothing resides on hard drive #2 that hasn't been lost already) or do whatever to fix this sort of problem?


Sorry for the awfully long post. I felt the need for detail (and the need to show how much I've been working on this).

Peter --
Find common answers using Google Groups:

http://groups.google.com/groups?group=comp.databases.ms-access

 

[wolluf]

http://support.wdc.com/download/index.asp

(I'd certainly try the zero fill first).

PS. Can we take it the problems began after the PM resize? And was that (only) on disk 2?

 

[FredWagner]

I would suggest that you might be overloading your system with IDE equipment. Also, you may be overheating things with too many drives with too little air circulation. Try leaving the covers off the cabinet, and directing a fan at the drives, see if they work better - then if they do, get a bigger case with more space between drives.
Also, IDE drives take up a LOT of CPU resources -I ran into a situation, loading Win2K on a server box with and IDE CD-ROM and using the SCSI controller on the motherboard, it would bomb out partway through the install. When I put in a separate SCSI controller in a PCI slot, the install went fine. Suggest you go with SCSI on your next PC, so the CPU doesn't get overtaxed trying to control too many devices at once. Fred Wagner
frwagne@longbeach.gov

 

[psemianonymous]

1. Thanks for the link. I will definitely do that tonight.

2. See, that's the problem. I have done a whole lot of retarded stuff on my computer. Sure, it happened sometime after the Partition Magic resize. (Details: Move a 50GB nearly full FAT32 partition forward 1.2GB, and shrink the partition accordingly) (More details: I wanted a DOS partition, and they have to start and a cylinder boundary or somesuch. So I decided the simplest way to ensure this is to add the DOS partition at the beginning of the drive.)
I'm just saying it could be partition magic's (version 8.0) fault. I'm not saying, though, that the problem occurred immediately after running PM resize. I actually got my RedHat 9 to install once fully (all the packages successfully copied) -- but then it halted the install at that time and I made an attempt to manually configure what was there. I'm no good at linux (I don't think anyone would enjoy configuring an entire system manually) so I gave up and attempted re-install. Note that this "nearly-complete" install was iteration#6 or in that area.

The other problem with the hard-drive#2-only theory is two things:
-Scandisk also failed after an unknown interval on scanning C drive (hard drive #1). Halted completely--no trace in scandisk.log
-My CD "verify" process failed when running the Red Hat CD verifyer
REBUTTAL: I had a swap partition on drive #2 at the time--could that affect the scan? Probably.


I still can't get around the fact that a scan of drive #1 failed. Today's scandisk is going over drive #2 again presently--it was still running after 5 hours at the time I left.



Peter --
Find common answers using Google Groups:

http://groups.google.com/groups?group=comp.databases.ms-access

 

[edfair]

Just from the description my stomach started churning with recollections of memory problems. Just a raw gut feeling with nothing concrete to go on. Ed Fair
Any advice I give is my best judgement based on my interpretation of the facts you supply. Help increase my knowledge by providing some feedback, good or bad, on any advice I have given.

 

[psemianonymous]

For tonight I have:
--Run utilities from Western Digital, see if anything obvious can be fixed. My grand hope is that the partition table overextended the drive size a little and thus fixing the MBR of drive#2 will fix everything.
--check overheating issues - Open case, run with fan. Possibly disconnect one drive for power issues. (probably the DVD).
--check/replace IDE cable. (suggestion from here at work) I have a backup, and this "sounds" right. This will take place after I have tried #1, #2.
--Lastly, check RAM. This is the "ickiest" of all, because I can't imagine my precious RAM going bad, and because I've had bad experiences (fried motherboard) with "failed" RAM installs. I really try to stay away from the internals as much as possible.


Anyway, this will obviously not be a one-night-only thing. I'll keep everyone posted on any progress I make. --
Find common answers using Google Groups:

http://groups.google.com/groups?group=comp.databases.ms-access

 

[BadBigBen]

Hi Foolio,

on the other thread what I meant was that there is a possibility of one of your drives going out (hardware failure) which can under some instances lock up the remaining drives, etc...

the reason why I am telling you this is that I had a secondary HD that would one day work perfectly fine and the next it disappeared and strangely my Master Drive had an extra 6GB (the size of my secondary Drive)... and all the scandisk etc. failed on me when this happened...

I also had a friends Comp that did strange things like you discribed and it boiled down to a failing CD Rom drive, after it was replaced all the problems disappeared...

Greets Ben

 

[firewolfrl]

Start burning your data to CDR disks.
work with one drive at a time.
after you are done backing up the drive do a low level format of the drive. (I would start with the 30gig first, then move data from 80 gig to 30 gig)
I suggest you use NTFS for any partition over 30 gig (works better)
upgrade your OS to an OS that works with NTFS
if you are using windows98 then keep the FAT32 partition size at 30 gig (stable)
download the upgraded Fdisk utility from MS (works with bigger drives and corrects the 30gig limit)
install the linux partitions at the end of the drive in a logical partition (let the install configure the partition, boot at the beginning of the logical partition where linux is installed)
do not let linux fool with the MBR
DOS at the beginning of the drive isn't necessary
BIGGER power supply will help.
you corrupted the partition table only a low level format will fix.

 

[psemianonymous]

Thanks firewolf. Short summary of events:

--I found a bad stick of RAM. Removed the RAM, and now everything is hunky-dory. Mostly.


--Linux still won't install. Red Hat error message still gives me some crap about there being invalid hard drive geometry, and it's Disk Druid program halts on several occasions (i.e. whenever I click on a FAT32 partition--instafreeze). I don't remember the error message exactly, and I keep forgetting to write it down. Anyway, I do believe that my Partition Magic'ing has pooped on my first hard drive's partition table. So I have done/am doing/will do all of your suggestions. I just finished backing everything up last night, and I plan to bring the machine to work tomorrow to let my co-workers try and figure it out--and let them do a reinstall/fdisk/everything.

I disagree about the DOS partition. I believe I *do* need it at the beginning of the first drive, for several reasons.

1. I need it to be on the first hard drive, because it has to boot off of the first hard drive. I use GRUB, and attempt to use the map (hd0) (hd1) command, but this doesn't work.
2. DOS has to boot at the beginning of a cylinder. I don't know how to do this except at the beginning of the drive.
3. Even if I did figure out how to do #2 properly, DOS still has to boot from a low address on the hard drive, below some limit like 2GB. I'm not concrete on this requirement, but I can guarantee that it hasn't worked SO FAR when attempting to load DOS at the end of my 20GB drive.


Also I'm steadfastly refusing to upgrade my Windows 98 for any reason. I might split hd#2 into two partitions, but I refuse to go to NTFS and a newer OS. Besides, a new OS costs money.

--
Find common answers using Google Groups:

http://groups.google.com/groups?group=comp.databases.ms-access

 

[psemianonymous]

Short recap:

1. FOUND PROBLEM: Bad RAM. I wasn't even looking, but sure enough, on computer start-up, it mis-counted my RAM on startup. Removed one RAM stick, and suddenly the computer "PASSES" my Red Hat 9 CD verification. Put in the other RAM stick, and they "FAIL". PROBLEM SOLVED.

2. PROBABLE SECOND PROBLEM: Partition Magic (...or one of the other 50 things I did...) messed my partition table. I am going to have to zero out the drive and start all the way over. This will take place Monday. I am confident that this is the only other problem. PROBLEM IDENTIFIED (not 100% sure).



I am extremely relieved that this is not an unfixable hardware problem. I was afraid my motherboard/IO Controller was broken.


Maybe I'll post again when I finally get linux installed. Until then, consider the problem SOLVED.

--
Find common answers using Google Groups:

http://groups.google.com/groups?group=comp.databases.ms-access

 

[firewolfrl]

I found that I can load a DOS fat32 partition anywhere on the drive as long as I follow these rules:
1. Work with ONE drive at a time.
2. Name the partition.
3. If you use Partition Magic to make the partition make sure you use a bootable floppy disk with the DOS program "Pqboot32.exe" or "Pqboot.exe" on it to make the partition the active bootable partition.
4. Always boot into the partition you just made. the boot process will give you an error because there is no Bootable commands installed.
5. Verify the partition by booting into it with a 98 startup floppy, use the "SYS C:" command to make the partition bootable

thats it.
if it does not boot to the DOS C:\ prompt you did not make the partition ACTIVE.

I use partition magic all the time. I have six 60 gig HDS on my system and I use the boot manager XOSL located at

http://www.xosl.org/

I always have 3 primary partitions and as many logical partitions as I want per drive.
on all six drives I load the XOSL and other DOS utilities as a small 56meg bootable partition as the third Primary partition. I do this as I use XP pro (NTFS) and find it is easier to configure as the first partition (for cloning purposes)
I have 98,XP,SusE Linux 8.0,Free BSD, and DOS loaded on the drives.
all the linux/unix partitions are loaded on logical partitions with the bootloader at the beginning of the logical partitions.
it took awhile to learn how to do it but it all works smooth now.