GORRIEDE@UREGINA1.BITNET (Dennis Robert Gorrie) (09/15/89)
Concerning difficulties with Micropolis 1674-7 drive and a C-LTD
SCSI 1000 controler:
One person, Josh Rovero, with an Amiga 500 and a Segate drive,
suggested to me that I edit my devsetup file, and change Flags BlockRead to
NoBlockRead. I did this, and then I tried to low-level format the drive,
using FEdX. FEdX does not have a MIC 1674-7 configuration, so I used the
1574-7 config file and modified it using FEdX. I only changed the
highcylinder to 1245, and I entered the manufactures disk error list.
Here is the format file:
/* --- Drive Configuration: Micropolis_1674-7 ----- Date: 09-10-89 --- */
Formatter CLtd-Format_1 /* Formatter ID */
ConFile Micropolis_0.fmt /* Configuration File */
CylCount 1245 /* Cylinder Count */
HeadCount 7 /* Head Count */
BlksPerTrk 36 /* Blocks Per Track */
BlockSize 512 /* Block Size */
Interleave 1 /* Interleave */
Patterns 3 /* Patterns */
BufferSize 250 /* BufferSize */
Comments: Micropolis 1674-07, Serial Number -> (mic9062910174 cs9132)
/* --- Bad Block List ------------------------------------------------ */
BadBlkFmt 1 /* Type 1: Cylinder, Head, Byte Offset */
BadBlk 8 5 15679
BadBlk 9 5 820
{ a list of 27 bad blocks from the manufacture spec sheet)
BadBlk 1061 3 18
BadBlk 1088 3 21
End BadBlk
___________________________________________________________________
I ran this format, and it seemed to work fine. It went through pattern
tests 0 thru to 6. Then, it started the "Final Format" and gave me this
error:
Beginning final format
Error - 26h - Invalid field in parameter list.
Program aborted - device error.
CLtd format failed - return code 100
So then I tried again, except I cleared the bad blocks list I had entered :
/* --- Drive Configuration: Mic_3 ----------------- Date: 09-11-89 --- */
Formatter CLtd-Format_1 /* Formatter ID */
ConFile Micropolis_0.fmt /* Configuration File */
CylCount 1245 /* Cylinder Count */
HeadCount 7 /* Head Count */
BlksPerTrk 36 /* Blocks Per Track */
BlockSize 512 /* Block Size */
Interleave 1 /* Interleave */
Patterns 7 /* Patterns */
BufferSize 250 /* BufferSize */
Comments: Micropolis 1574-07, Serial Number -> ()
/* --- Bad Block List ------------------------------------------------ */
BadBlkFmt 2 /* Unspecified Format */
End BadBlk
______________________________________________________________________
This time I ran the format it completed without errors. When it completed
all 7 pattern tests though, it had not found any bad blocks.
Questions:
Does my bad block list cause a syntax error when the format program tries
to run it? How am I to enter my bad blocks list? How safe is the drive to
use without this list of bad blocks not marked off? (Or has the drive
already marked them off, at the manufactures test site?)
Next I AmigaDOS formatted the drive using the same mountlist as above:
>system/format drive dh0: name Test1 ffs
The format stopped at Cyl 1239.
Question : Why did it not format all the way to 1245? On a MAC the same
drives go right to 1245, but on AmigaDOS they stop short, or so I am told
by the supplier.
I changed the mountlist to have HighCyl at 1235. I rebooted, and AmigaDOS
formated the drive again. This time it reached 1235, and finished ok.
I made a second partition , from 1236 to 1245. I tried several times to
AmigaDOS format this partion, but it would not even do the first cylinder.
I then copied my workbench software to the drive, and got it working.
I tried to use the "Find Errors" program that comes with SCSIDOS 3.0.
Every time I tried to run the program, it would scan a few tracks on the
hard disk, and then hang. The red drive light would go out. Nothing I
typed would get the program to exit. DO you have any ideas why it keeps
crashing after a while? (C-LTD techician says I don't have enough
chip memory.)
As a test, I edited the devsetup file and changed NoBlockRead to
BlockRead. I used the devspeed program to compare the difference:
(I changed the devsetup file and warm-booted between tests)
flags NoBlockRead
6 K Reads 145221 Writes 133120 b/sec
15 K Reads 184320 Writes 153600 b/sec
51 K Reads 208271 Writes 198193 b/sec
153 K Reads 205944 Writes 199264 b/sec
307 K Reads 209454 Writes 201442 b/sec
512 K Reads 211862 Writes 203443 b/sec
flags BlockRead
6 K Reads 319488 Writes 99840 b/sec
15 K Reads 335127 Writes 153600 b/sec
51 K Reads 438857 Writes 189046 b/sec
153 K Reads 460800 Writes 196085 b/sec
307 K Reads 460800 Writes 200347 b/sec
512 K Reads 472615 Writes 202772 b/sec
I left Flags BlockRead in the devsetup file to see if it would now work
without errors.
I made my own AmigaDOS "script" to fill up the disk to try and test it:
>makedir t
>copy #? t all
>cd t
>execute script
I let this run until the drive was full (155 megs) I then made a similar
script file to delete everything. I ran this once, and did not get any
errors.
1) The previous technician at C-LTD suspected that the drive was damaged.
How can I tell if it is damaged?
2) Assuming the drive is NOT damaged:
The FEdX program and formatter will not process my bad-blocks list. Can
FedX be fixed, or can I get another formatter that will process my bad
blocks list?
3) The "Find Errors" program always crashes. How can I test the data
safety and integrety of this drive?
4) The NoBlockRead flag in the devsetup file seemed to fix my low-level
format, to SOME degree. However, the drive now seems to run fine with the
BlockRead flag enabled. How safe is it to run the drive with BlockRead
enabled, and is it worth the speed increase?
5) Why did the AmigaDOS format quit at cyl 1239 instead of 1245? When I
make a partition of 1236 to 1245, why will it not even format one
cylinder?
6) Is the KRONUS controller available for Amiga 1000s, and would it solve
any of the above problems? Will an upgrade or trade-in be offered?
I realize I am asking a lot. I suspect the suppliers use a different
computer or software to do the low-level format. It seems quite clear they
do NOT use the FEdX and Formatter programs.
MANY THANKS again to Josh Rovero for his hint in getting the drive to work.
It saved me a lot of grief.
||| Dennis Gorrie ||| GORRIEDE AT UREGINA1.BITNET ||||