Amiga.org
Amiga computer related discussion => Amiga Hardware Issues and discussion => Topic started by: Im>bE on July 22, 2003, 02:11:49 PM
-
story:
I had set it up
and was copying from my old hd.
but I used an old program called diskmaster
and when it reached the 4 gig limit,
diskmaster damaged the rigid disk block
and no partition on it worked after that.
problem:
then I tried to reinstall it with hdtoolbox,
but found that when I 'read configuration'
it gets incorrect values and incorrect size.
I tried to find the info on all these values
on the net,
but there was 2 fields that I never found.
those were:
'blocks per track'
'blocks per cylinder'
MadMatty told me that hdtoolbox
has its own drive geometry
so that it creates these 2 fields itself,
so if you have an
80 gigabyte diamondmax plus 9,
model name:6Y080L0
and has installed it with hdtoolbox
then PLEASE tell me what these 2 field have.
or tell me what I can do to fix it an other way...
or else I may have to give it up and throw it out the window. :(
-
What "incorrect" values does it give? How do
you know they are incorrect?
I'd be more willing to trust what HDToolBox
says. If you continue, does it work anyway?
You shouldn't attempt to fill in the various
boxes with values as most are not directly
related to the disk in modern units.
-
You need os 3.5 onwards to use an 80 gig drive. :-?
-
thanx for replying.
I use 3.9, but at the time I used diskmaster
I might have lacked the smartfilesystem
handler in l: tho the sfs partition worked.
so im not sure if it could have helped the hd crash.
>What "incorrect" values does it give?
Cylinders=8127
-supposed to be 16383
heads=8
-supposed to be 16
blocks per track=254
-have no idea what it should be.
blocks per cylinder=2032
-no idea
park head where=8127
-supposed to be 16383
hey! seems like all values could have been halved.
how wierd...
>How do you know they are incorrect?
technical specs of the exact hd
on various net sites.
>I'd be more willing to trust what HDToolBox
>says. If you continue, does it work anyway?
yes, but with much less space. (7.8 gig)
>You shouldn't attempt to fill in the various
>boxes with values as most are not directly
>related to the disk in modern units.
oh.
the only value that is correct,
is 'bytes per blocks'=512
and it can't be changed either.
I think im gonna try double all values.
might be interesting.
-
Im>bE
don't worry about the wrong value in HDToolbox.
This kind of problems, are well known with the original IDE port (A1200 and A4000).
If HDToolbox show <8GB of total space, could be a Filesystem problem.
Have you chose and selected SFS Filasystem from HDTollbox? (SmartFileSystem must be in L:)
Ciao
PS- have you installed BB-1 and BB-2 on your OS3.9?
-
>Have you chose and selected SFS Filasystem from >HDTollbox? (SmartFileSystem must be in L:)
well.. this is not very relevant
until the hd is installed.
I don't think it changes anything
if I go to the partition screen
and install sfs BEFORE installing the hd.
>PS- have you installed BB-1 and BB-2 on your >OS3.9?
yes.
I guess I should forfeit
all the time consuming copying I did,
and just manually adjust the 'blocks per track'
field to give about 75 gigs (the actual size)
and then start over. :(
DAMN I regret I didn't note down the RDB
or took a snapshot of it... :(
unless someone can help me
within the next... hour.
-
now I have reinstalled the hd
with 2 partitions inside the 4gig limit,
and 1 gigantic partition streaking from inside the 4gig limit to somewhere close to the end.
I copied everything from my old hd
(3.8gig) without rebooting,
and when I reboot,
I find that the gigantic partition is
'unitialized'...
This was not expected.
but it happened anyway,
and im going slightly insane
from all the time I have spent trying to
get this hd set up and filled with all my data.
the gigantic partition uses smart filesystem
wich is supposed to work ok beyond 4 gig,
but Lempkee told me that there could occur
a problem with buffers when copying lots of data,
wich is what I did.
but if sfs can't handle lots of data,
then what filesystem should I use instead,
or what can I do to fix this partition?...
or where do they sell strong ropes? B)
-
Hmm.. I have one 40G hd as one big partition with SFS and it works fine. Although I had to disable the automount from it and mount the partition in setpatch with latemount in devs:nsdpatch.cfg. Otherwise there comes problems in cold boot, because setpatch (or blizkick) hasn't updated the scsi.device yet. I don't know if you've tried that too or could it help...
I never have had problems copying lots of stuff with sfs... if there isn't bugs in newer SFS's. I've been using 1.84.
PFS3 is another option. If you use DirectSCSI version of it, then there shouldn't be any problems with anything else. Doen't matter which scsi.device you have etc.
And you have probably double checked the scsi.device too :) That it's updated for sure.. ie not disabled with some setpatch option or missing amigaos rom update file in devs or similar :)
-
@ Im>bE
with large and modern HD, the boot partition MUST be the first on the HD and less than 2 GB.
Ciao
-
According to http://www.maxtor.com/en/products/ata/desktop/diamondmax_plus_9/model_numbers_table.htm the data of your hdd should be as follows:
heads: 16 (= tracks per cylinder)
sectors: 63 (= blocks per track)
cylinders: 158,816
blocks per cylinder: 1,008 (= blocks per track * tracks per cylinder)
total number of blocks: 160,086,528 (heads * sectors * cylinders)
bytes per block: 512 (assumed by me)
hdd size in bytes: 81,964,302,336 (number of blocks * block size)
hdd size in gigabytes: 76.34 (bytes/1024/1024/1024)
Now, the only value that really counts is the total number of blocks. The latest version of HDToolbox suggests ca. 1MB per cylinder. The first two cylinders are used for the RDB and it must fit there.
So the number of blocks per cyl should be 2,016 and the number of cylinders 79,408.
Bye,
Thomas
-
Oh, and of course the rules:
if you connect the hdd to the internal IDE bus:
- the boot partition must completely reside below the 4GB limit.
- on the boot partition the required patches must be installed (SetPatch, NSDPatch.cfg, AmigaOS ROM Update)
- you must use a filesystem that supports NSD commands (FFS V44+ or SFS)
For PFS3 you need IDEfix97. It is not true that PFS3ds does not give any problems. The internal scsi.device IDE driver has an 8GB limit for SCSI-Direct commands.
If the hdd is connectred to a Phase5/DCE device (BlizzardPPC, CyberstormPPC etc.) and you use either PFS3, PFS3ds or SFS, there is no limit in boot partition size. You can even create one big 80GB boot partition. For FFS still NSDPatch is needed and thus the boot partition must be below 4GB.
If the hdd is connected to another controller, its rules apply. They can be different or similar, depends on each single controller.
To check if all works as it should (before copying any files) you can download the Check4GB archive and run the program CheckHD.
http://uk.aminet.net/aminetbin/find?check4gb
Bye,
Thomas
-
You're trying to solve the wrong problem. HDToolBox uses its own algoritm to decide sensible values for everything, as long as it all adds up to the same total number of bytes (more or less ;-) ) there's NO reason why you need 16 heads instead of 8 on a harddrive that only has 4 ACTUAL heads inside it anyway. Kapiche?
The problem you have is that your FFS and scsi.device doesn't support large disks. Make sure you've got the LATEST FFS on disk and inside the RDB, and make sure setpach is run in the startup-sequence. WITH the scsi.device patch.
-
I followed Thomas calculations blindly,
and my hd has those values now.
ppl informed me that it was all about
the scsi.device as olegil points out.
then I tried to get the new scsi.device
into the softkicked rom.
without luck...
(my os3.9 is messed up)
After I was at the bottom,
after many a frustration,
jPV suggested that I should use direct scsi.
I did, and it works. (for now)
so I am happy again.
thank you for helping lame me.
ps: I can't hear the hd
unless it reads/writes from it.
damn, its quiet...
(maybe because I have it on a soft pillow)
I think I would reccomend this hd
for its pure antinoise.
-
Now the scsi direct partition can't write
beyond 7.8 gig of the hd... :(
I use pfs3ds filesystem.
(professional filesystem 3 direct scsi)
When I try to copy a file to the partition
so that it exceeds the 7.8 limit,
then I get a 'block id error 5' error message
-REREAD|CANCEL-
and when I try to reread,
it just skips a heap of blocks
and I get the same error.
when I eventually cancel
I get a 'hd update failed'
and all files and dirs are not listed
until I reboot.
Ive never heard of anything similar before.
-
by Im>bE on 2003/7/24 20:05:00
Now the scsi direct partition can't write
beyond 7.8 gig of the hd... :(
AFAIK with the standard IDE interface and PfsDirectScsi, the limit is near 8GB.
I think you need IDEFix97 with plain PFS3.
If you'll install IdeFix97, you must add the option:
SetPatch SKIPROMUPDATES "scsi.device" QUIET
Ciao
PS- IMHO buy a FastATA interface (PowerFlyer PioMode4) :-)
-edit
have you checked that the old HD partitions, has the same Mask and MaxTransfer like the new one?
-
He doesn't have the os rom updates in use at all, so the setpatch option isn't needed :) If he'd have the updated scsi.device, then there wouldn't be any problems at all :)
-
>If he'd have the updated scsi.device, then there wouldn't be any problems at all :)
ye, tell me about it.
the thing is that the rom updates
either crash or slow down reading (not writing)
on any of my hds.
im trying to use blizkick to install the scsi.device atm.
-
no luck with blizkick.
it says: "mapped rom, but unable to plant module"
when I try to plant the scsi.device module.
now I have enabled the rom update again,
to see if it is any better,
but it is still slow as hell when reading from hd.
I would say about 1/4 speed as without the update.
WHY is it slowing it down so much???
is this ONLY to me?
is there something wrong with the scsi
Is there something wrong with my miggy?
I just can't get myself to use this slow slow disk access.
Its true! reading from hd gets darn slow
with the BB39-2 rom update on my A1260 aga.
I also think the rest of the system gets equally slower...
-
by Im>bE on 2003/7/25 0:59:23
>If he'd have the updated scsi.device, then there wouldn't be any problems at all :)
ye, tell me about it.
the thing is that the rom updates
either crash or slow down reading (not writing)
on any of my hds.
im trying to use blizkick to install the scsi.device atm.
But . . . . you haven't the RomUpdate running???
You'll never see all the 80GB of your new HD !
So, use PFS with Idefix97.
Ciao
PS-for the speed problem, try to disable ONLY the RomUpdate scsi.device with:
SetPatch SKIPROMUPDATES "scsi.device" QUIET
-
>So, use PFS with Idefix97.
I hate idefix.
it fxxks up EVERY possible thing it can.
>PS-for the speed problem, try to disable ONLY the RomUpdate scsi.device
I could do that,
but the new scsi.device is the ONE thing I want,
so if I find out that the rom update runs perfectly
WITHOUT it, then I might
blow up my head in PURE irritation!
any other device that could slow down my system?
-
My head just blew up.
Im scraping off the mess off my monitor as I type...
so when the -only- thing I need slows down reading from hd to 1/4, what do I do?
is there anything I can do
to fix the new scsi.device to work on my setup?
nah, probably not...