[NBLUG/talk] HDD problems

[Robert Hayes]

I've just moved a spare HDD from an older machine to a newer one.
I had some trouble mounting it, as per earlier posts.

Now I've emptied the drive of important data, and have mkfs'd an ext3 
partition on it. Still having problems that I can't unravel, tho:

The drive is a Maxtor 10GB 5400 RPM unit that was factory issue with my Dell 
800 MHz machine a few years ago. It's now in my 2.2 GHz desktop. The OS is 
Debian 3.0.

The current machine bios sees the drive as the secondary slave.

When I ran cfdisk and made the drive one partition, 10GB, Type 83, everything 
was fine. Then I ran cfdisk again to check the settings, and received the 
following: 

No partition table or unknown signature on partition table
Do you wish to start with a zero table [y/N] ?

If I run fdisk, this results:

Device contains neither a valid DOS partition table, nor Sun, SGI or OSF 
disklabel
Building a new DOS disklabel. Changes will remain in memory only,
until you decide to write them. After that, of course, the previous
content won't be recoverable.

The number of cylinders for this disk is set to 1215.
There is nothing wrong with that, but this is larger than 1024,
and could in certain setups cause problems with:
1) software that runs at boot time (e.g., old versions of LILO)
2) booting and partitioning software from other OSs
   (e.g., DOS FDISK, OS/2 FDISK)
Warning: invalid flag 0x0000 of partition table 4 will be corrected by w(rite)

I'm able to mount the drive manually, but of the 10GB it only reveals 8.67GB 
claiming 33MB are used!

Also, everything going to or from the drive is very slow. The cpu cranks up to 
100%, and file copy or move operations stall. They eventually finish, but it 
is unacceptable.

I've tweaked with hdparm and that yields the following. Turning on dma makes 
no difference on the speeds.

/dev/hdd:
 multcount    = 16 (on)
 IO_support   =  3 (32-bit w/sync)
 unmaskirq    =  1 (on)
 using_dma    =  0 (off)
 keepsettings =  1 (on)
 readonly     =  0 (off)
 readahead    =  8 (on)
 geometry     = 1215/255/63, sectors = 10000000000, start = 0

/dev/hdd:
 Timing cached reads:   1272 MB in  2.00 seconds = 636.00 MB/sec
 Timing buffered disk reads:   14 MB in  3.06 seconds =   4.58 MB/sec

The 30GB/7200RPM drive in this machine has the same cached read rates, but the 
buffered reads are ten times faster: ~45.8MB/sec.

fdisk -l /dev/hdd results:

Disk /dev/hdd: 10.0 GB, 10000000000 bytes
255 heads, 63 sectors/track, 1215 cylinders
Units = cylinders of 16065 * 512 = 8225280 bytes

Disk /dev/hdd doesn't contain a valid partition table

I've also gone through and done a mkfs.ext with the -c -c options to 
write/read check for bad blocks. It seems to be clean.

Any ideas, folks?

I'm either very thick or very stumped.
Or a thick stump.

Thank you,

Robert