OS/2 Shareware BBS: 36 Tips/36-Tips.zip/cy1024.zip/CY1024.TXT

Text File | 1992-08-14 | 22KB | 509 lines

#: 11808 S4/General 2.x Q&A
    29-Jul-92  11:55:37
Sb: OS/2 2.0 & SCSI 1.5 Gb
Fm: Andrew Winch 71740,3224
To: Support for OS/2

I have a .5 gigabyte hard disk with version 2.0 in it, working fine.
Same version under a 1.5 disk fails drastically. It blows the disk device
drives (scsi drivers).  OS/2 will not boot with such drive.  Is the a core
limitation?   Thanks.

* Replies:  11852, 12065, 12078

#: 11852 S4/General 2.x Q&A
    29-Jul-92  20:16:24
Sb: #11808-OS/2 2.0 & SCSI 1.5 Gb
Fm: Mercer Harz 70431,150
To: Andrew Winch 71740,3224 (X)

First, I'm not an IBMer.  Just an OS/2 enthusiast.  For the best support, you
should post your question in Section 10, Hardware, in the IBMOS2 forum, as
well as download the file PROBLM.TXT in library 17 of IBMOS2, fill it out and
mail it in (via CIS email).  However, you are here, now, so I will try to
respond.

There is no core limitation that I'm aware of.  As far as I know OS/2 should
be able to address volumes at least 2GB in size.  There are, however, limits
imposed by some hardware, and the device driver being used may need updating,
changing or removal.  Are your two systems identical in every respect except
disk capacity?  To help you better, we need to know more about your systems,
such as the make/model of the drives, the disk controllers, the system
units/motherboards, installed RAM, video type, device drivers, etc.  Since
this is essentially the same info that you need in order to fill out the
Problem Report Form in the PROBLM.TXT file, you probably should download it,
fill it out, and send it to the CompuServe address that is shown on the form.
Then, repost your question and give us just a short synopsis of your systems'
configurations.

Thanks,

   Mercer

#: 12065 S4/General 2.x Q&A
    31-Jul-92  07:30:35
Sb: #11808-OS/2 2.0 & SCSI 1.5 Gb
Fm: Bernd Kunrath 100015,2526
To: Andrew Winch 71740,3224

Hi Andrew, no there's not a limitation in OS/2 2.0. We have two servers
running with large disks. One with two 1.53GB Micropolis on a Future Domain
TMC850 SCSI adapter, one with one 1.53GB Micropolis on a Adaptec 1542B. Both
are working fine with standard 2.0 (no fixes) and IBM LAN Server 2.0. Tell me
your setup to see if i can help. Ciao, Bernd

#: 12078 S4/General 2.x Q&A
    31-Jul-92  08:57:11
Sb: #11808-OS/2 2.0 & SCSI 1.5 Gb
Fm: Kamran Rejaee [IBM] 70733,1401
To: Andrew Winch 71740,3224

Hi Andrew

  Obviousely your are dealing with a hard disk that has greater than 1024
cylinders. I am quite sure that 2.0 does support such geometry, however I
would appriciate it if you were to give me the type and brand of your hard
disk controller plus your hard disk geometry. I suspect the problem to be with
your controller and the device driver your are using with it.


                      Thanks
                             Kamran [IBM]

* Reply:  12174

#: 12174 S4/General 2.x Q&A
    31-Jul-92  16:40:22
Sb: #12078-OS/2 2.0 & SCSI 1.5 Gb
Fm: Wayne Kovsky [KSW] 76164,3504
To: Kamran Rejaee [IBM] 70733,1401

Kamran, if OS/2 2.0 does support a hard disk that has more than 1024
cylinders, there are many of us (certainly, I am one of them) who are not
aware of it.  I have an ESDI drive with more than 1024 cylinders, and it is
supported only because the controller can remap the drive so that it appears
to OS/2 that the drive has fewer than 1024 cylinders.  Other people have
drives that cannot be remapped, and they are literally letting a substantial
portion of their disks go unused, because of this limitation.

When you find definitive information, please do post it here in a very
conspicuous manner.  OS/2's non-support of drives with more than 1024
cylinders is a sore point with many people - it would be a terrible shame if
we were all victims of mis-information!

* Replies:  12399, 12406, 12696

#: 12399 S4/General 2.x Q&A
    01-Aug-92  23:03:28
Sb: #12174-#OS/2 2.0 & SCSI 1.5 Gb
Fm: Kevin O'Connor 73737,2345
To: Wayne Kovsky [KSW] 76164,3504 (X)

>I have an ESDI drive with more than 1024 cylinders, and it is supported
>only because the controller can remap the drive so that it appears to OS/2
>that the drive has fewer than 1024 cylinders.

Which controller are you using? I have an UltraStor 12F running a >1024
cylinder Micropolis ESDI driver WITHOUT remapping. I have all 600+ megs of
disk space available. The only hard disk driver I have in my config.sys is
IBM1S506.ADD.

Under OS/2 1.x this requires a driver from UltraStor, but it works without
the driver under 2.0.

* Reply:  12501

#: 12501 S4/General 2.x Q&A
    02-Aug-92  14:41:50
Sb: #12399-#OS/2 2.0 & SCSI 1.5 Gb
Fm: Wayne Kovsky [KSW] 76164,3504
To: Kevin O'Connor 73737,2345

I'm using a DTC 6280 controller, Kevin.  If I can use this controller without
remapping, it is news to me.  As I said in my prior post, it would be a shame
if many of us are either remapping or just giving up on getting full use of
our disks, through being misinformed.

By doing a low-level format of the disk into 60-sector mode, I get the full
324MB capacity of my drive.  The controller also supports a 17-sector mode for
Novell, in addition to "native" mode, both of which result in more than 1024
cylinders.

I did a low-level reformat several times during Beta testing, for two reasons.
One was to see if the 1024-cylinder limit had yet been removed (it had not,
each time I tried it).  The second reason was that high-level formatting was
unreliable under Beta versions, and bad tracks that were marked at the ESDI
level were not mapped out during the high-level formatting (this was not fixed
until GA).  I had to low-level-format my drive in a mode that maps bad sectors
in a cylinder to a spare sector reserved for that purpose (which costs
available disk space).

The last low-level format I did was just prior to installing GA, but I did not
try the 1024-cylinder limit at that time - it is entirely possible that it is
removed now, and I just did not know it.  I have seen a number of posts both
here and on FidoNet from people who are not able to remap their drives at all,
and are thus just not able to use the full capacity of their disks.  I hope
someone from IBM will confirm that this limitation is gone, and then publicize
it widely - it is the cause of a lot of unhappiness and criticism right now.

* Reply:  12697

#: 12697 S4/General 2.x Q&A
    03-Aug-92  13:56:10
Sb: #12501-#OS/2 2.0 & SCSI 1.5 Gb
Fm: Kamran Rejaee [IBM] 70733,1401
To: Wayne Kovsky [KSW] 76164,3504 (X)

You should be able to use you large media, I recommend reformatting your drive
again and see what happens. Make sure that your C: drive is not larger than
1.0 gig. I posted a note early today in regards to this problem and I hope
that others concerned with this prob. will see it.
LEt me know if I can help you anything Thanks

* Reply:  12734

#: 12734 S4/General 2.x Q&A
    03-Aug-92  15:57:57
Sb: #12697-OS/2 2.0 & SCSI 1.5 Gb
Fm: Wayne Kovsky [KSW] 76164,3504
To: Kamran Rejaee [IBM] 70733,1401

Thanks for the response, Kamran, but "reformat your drive and see what
happens" is not my idea of a good time <g>!  I did that repeatedly during
Beta, because IBM repeatedly said "the bad-spots problem is gone - maybe -
reformat your drive and see what happens".  (It was fixed in GA code for the
first time.)

Anyway, backing up 320MB of data (without a tape drive, since my tape does not
have an HPFS-aware driver), reformatting, and then trying to use the drive
just to "see what happens" isn't very interesting to me.

I <did> reformat the drive for every Beta release, and I <did> try to use
standard-mode low-level formatting each time; it never worked.  I don't recall
if I tried this when I got GA code - it is likely that I did not - but if this
works now, then it works in GA code for the first time.

Now, if you tell me that this was fixed just in time for GA code, and sure
enough, it didn't work until then, I <might> try reformatting my drive just to
see if it works.  But, if you think it worked before GA code, I can tell you
that it did not, and I don't want to waste my time on the off chance that it
might work now.

I'm not trying to be difficult, Kamran - I really do appreciate your response.
 However, other people have installed 2.0 GA code as their first-ever OS/2
system, and they have reported this problem (although with different
controllers than I have).  I need a more confident assertion from you ("This
<does> work with a DTC 6280 ESDI controller like you have") before I undertake
such a big step.

#: 12406 S4/General 2.x Q&A
    01-Aug-92  23:46:05
Sb: #12174-#OS/2 2.0 & SCSI 1.5 Gb
Fm: John Bridges [GRASP] 75300,2137
To: Wayne Kovsky [KSW] 76164,3504 (X)

Wayne, there was a discussion about this some time ago in IBMOS2. OS/2 does
support more than 1024 cylinders, in fact I am using it with a 650meg ESDI
drive right now without any remapping.

The main problem is PMFDISK which does not support more than 1024 cylinders. I
used SpeedStore to setup the partitions, then formatted the first one for FAT
and the second one for HPFS using the normal OS/2 format routine.

* Reply:  12502

#: 12502 S4/General 2.x Q&A
    02-Aug-92  14:41:58
Sb: #12406-#OS/2 2.0 & SCSI 1.5 Gb
Fm: Wayne Kovsky [KSW] 76164,3504
To: John Bridges [GRASP] 75300,2137

Well, that is new information, John (see my reply to Kevin O'Connor).  I knew
that I could not use >1024 cylinders on any version prior to GA, but I
couldn't swear that I ever tried it for GA code, so I thought I had simply
missed a key development.

If FDISKPM (and presumably, FDISK) don't support partitioning a disk with more
than 1024 cylinders, then I still can't use it.  Although I have SpeedStore
and several other low-level disk utilities, none of them would support the
MultiBoot feature (correct me if I am mistaken on this point).  While I could
give up this facility more easily now than I could have during Beta, I still
need it.

I have a solution that works for me:  I have access to the full capacity of my
hard drive, and I am giving up nothing in performance or reliability for that
access.  However, that is not true for a number of other people.  If the
limitation is in FDISK/FDISKPM, then it is still a bug, and it should be
fixed.  (And, for people who own SpeedStore or other such utilities, it should
be made clearer that this option is available to them, perhaps with a note in
the READ.ME file that comes with OS/2.)

* Replies:  12678, 12841

#: 12678 S4/General 2.x Q&A
    03-Aug-92  10:52:57
Sb: #12502-OS/2 2.0 & SCSI 1.5 Gb
Fm: System Integrators 70750,1573
To: Wayne Kovsky [KSW] 76164,3504 (X)

It is my belief that the problem with the 1024 cylinder limit is a factor of
the BIOS, not OS/2. Once OS/2 has loaded its support, there is no inherit
problem with drives greater than 1024 cylinders.

The problem is that BIOS has a 10 bit cylinder address. This leaves address 0
... 1023.

I would believe that if you could create the partition within the first 1024
cylinders; place the OS2LDR, OS2BOOT, OS2KRNL, and the support files for the
SCSI drivers within the first 1024 cylinders then OS/2 should boot and run
just fine.

#: 12841 S4/General 2.x Q&A
    03-Aug-92  22:46:27
Sb: #12502-OS/2 2.0 & SCSI 1.5 Gb
Fm: John Bridges [GRASP] 75300,2137
To: Wayne Kovsky [KSW] 76164,3504

Wayne, here are the problems:

The normal BIOS information on number of tracks is only 10 bits. Even though
some BIOSs support more than 1024 cylinders. FDISKPM does not, it only see's
the 10 bits (at least on ESDI drives).

The partition sector is also designed for 10 bits for the starting cylinder.
But OS/2 does not do an checking to see if the starting cylinder plus the
number of cylinders adds up to more than the total number of cylinders
specified by the BIOS.

That's how you can access more than 1024 cylinders under OS/2.

(FORMAT works fine with more than 1024 cylinders, the only problem is
FDISK/PMFDISK).

 If you have a copy of the Norton utilities you can edit your partition table
quite easily.

* Reply:  12954

#: 12954 S4/General 2.x Q&A
    04-Aug-92  11:27:13
Sb: #12841-OS/2 2.0 & SCSI 1.5 Gb
Fm: Wayne Kovsky [KSW] 76164,3504
To: John Bridges [GRASP] 75300,2137

John, I presume by "BIOS" you mean the disk controller BIOS, and not the
system BIOS?  If not, then I still don't understand the BIOS limitation, since
OS/2 replaces the ROM BIOS during boot.  If my boot manager partition were
within the first 1024 cylinders, and the OS/2 2.0 partition I wanted to boot
from were also within the first 1024 cylinders, then the ROM BIOS should not
be a factor.  As soon as the boot sector record is read and the first pieces
of OS/2 are brought into memory, OS/2 should from that point forward be able
to deal with more than 1024 cylinders.  Prior to that point, the ROM BIOS
would not be asked to deal with more than 1024 cylinders, so it should be
"happy".

However, if you <did> mean the disk controller BIOS, then it seems to me that
it is clearly not a problem.  My disk controller can deal with much greater
than 1024 cylinders already, in any of three modes (normal mode, Novell mode,
and 60-sector translation mode).

I don't have Norton, but I do have PC Tools, HDTest/HDFormat, and several
others.  I used to do a lot of mucking about with the partition table, but I
would be reluctant to do that now (with the added complexity of Boot Manager,
for one reason).  If the "bit surgery" is rather simple, and you can post a
cookbook for which bits must be changed to what new values, please do.
However, if it requires any actual thought, I probably won't attempt it!

* Reply:  13071

#: 13071 S4/General 2.x Q&A
    04-Aug-92  18:27:24
Sb: #12954-#OS/2 2.0 & SCSI 1.5 Gb
Fm: John Bridges [GRASP] 75300,2137
To: Wayne Kovsky [KSW] 76164,3504 (X)

Wayne, In case I wasn't clear, by the BIOS I was refering to the information
the BIOS provides on the drive size, PMFDISK uses this information.

 I know OS/2 uses it's own code to reference the device, this is in fact why
OS/2 works with more than 1024 cylinders (since the controller is not limited
to 1024 cylinders, but the BIOS call to read/write data from the drive uses 10
bits for the cylinder number, and is hence limited to 1024 cylinders).

Using Norton DE (DiskEdit), to edit the partition section is really not very
hard since it displays the data in a easy to read format. Just type in the new
numbers and away you go.

Or maybe IBM could add a switch to PMFDISK that lets you specify the true
number of tracks on the drive??  Or maybe they could fix PMFDISK so that it
recognizes extended drive size information (as most new BIOS's support (like
AMI)).

* Reply:  13124

#: 13124 S4/General 2.x Q&A
    04-Aug-92  22:18:23
Sb: #13071-OS/2 2.0 & SCSI 1.5 Gb
Fm: Wayne Kovsky [KSW] 76164,3504
To: John Bridges [GRASP] 75300,2137

Thanks for the info, John.  I liked your last idea best:  "Or maybe they could
fix PMFDISK so that it recognizes extended drive size information".  I think
I'll wait for the fix.

* Replies:  13157, 13181

#: 13157 S4/General 2.x Q&A
    05-Aug-92  01:34:14
Sb: #13124-OS/2 2.0 & SCSI 1.5 Gb
Fm: John Bridges [GRASP] 75300,2137
To: Wayne Kovsky [KSW] 76164,3504 (X)

Wayne, I wouldn't hold your breath for a fix for PMFDISK, because normal DOS
and FDISK do not handle more than 1024 cylinders either (without special
drivers).

I have to load SSTOR.SYS in my config.sys to get DOS to use my second
partition correctly (or else it wraps around at the 1024'th cylinder (which is
VERY NASTY)).

Even considering the problems with PMFDISK, I was very excited to see that
OS/2 would use my second partition (which goes beyond 1024 cylinders) without
any special drivers, patches or fixes!

#: 13181 S4/General 2.x Q&A
    05-Aug-92  07:40:15
Sb: #13124-OS/2 2.0 & SCSI 1.5 Gb
Fm: Bill Hinkle 75300,2050
To: Wayne Kovsky [KSW] 76164,3504 (X)

Wayne, I recall Mel Hallerman making mention of the extended cylinder count
APIs some clone BIOSes.  He basically said that they had looked into
supporting that capability and found that it was not entirely reliable or
consistent.  He didn't mention any details; I suppose just being able to
figure out, reliably, whether there is such support or not could be a problem.
Maybe he'd expand on that. <<Bill>>

#: 12696 S4/General 2.x Q&A
    03-Aug-92  13:44:01
Sb: #12174-#OS/2 2.0 & SCSI 1.5 Gb
Fm: Kamran Rejaee [IBM] 70733,1401
To: Wayne Kovsky [KSW] 76164,3504 (X)

Guys...
OS/2 2.0 definitely does support large media (i.e. CYL > 1024). I know it
because I was involved with its development. Now if there is problem with
particular HD controller, by which you cant access your full HD capacity, I
recommend cantacting the manufacturer of the device. They might be able to
provide you with the appropriate driver. I know that SCSI controllers such as
ADAPTEK 1642b, 1542 and etc.. work just fine since I got one myself. If your
system persists on not working, I suggest contacting the support line and
reporting it as a deffect. Meanwhile I'll be glad to help out if you give me
more detailed info.


          thanks     Kamran [IBM]

* Reply:  12735

#: 12735 S4/General 2.x Q&A
    03-Aug-92  15:58:07
Sb: #12696-OS/2 2.0 & SCSI 1.5 Gb
Fm: Wayne Kovsky [KSW] 76164,3504
To: Kamran Rejaee [IBM] 70733,1401

Kamran, my controller is a DTC 6280 ESDI controller, and I have contacted DTC
to see if they were aware of any problems; they were not.  Since this is not a
SCSI controller, there is no way for them to provide me with "the appropriate
driver".

I'm not sure what detailed info you need, but if you can be more specific, I
will try to provide it.

I wonder if this isn't a problem of <some> of the software supporting this
properly, and other parts of the software not doing so.  Someone said that
FIDISK/FDISKPM do not support more than 1024 cylinders - perhaps it does for
SCSI drives but not ESDI drives.  Or perhaps FORMAT (which was badly broken
right up until GA code) does not properly support more than 1024 cylinders, at
least for ESDI drives.

I accept that SCSI works, since both you and others have reported success with
it.  But are any lurkers using an ESDI drive with more than 1024 cylinders,
without having to low-level format the drive in a way that "hides" the actual
number of cylinders from the OS?

* Reply:  12919

#: 12919 S4/General 2.x Q&A
    04-Aug-92  09:20:49
Sb: #12735-#OS/2 2.0 & SCSI 1.5 Gb
Fm: Kamran Rejaee [IBM] 70733,1401
To: Wayne Kovsky [KSW] 76164,3504 (X)

Wayne, I am trying to find out if there is a problem with your controller,
that should'nt be too long. I am not aware of any problems with fdisk and
fdiskpm, off course I could be wrong! right?
In your master boot record (MBR) and Extended Boot Record (EBR) there two 32
bit fields known as RBA. When you use fdisk to make a partition, the relative
sector of your new partition to the MBR and its length are recoded in the RBAs
in sectors units. Now if the initial number of Cylinders reported to the fdisk
are erroneous, naturally the disk size will be incorrect. For example format
will not update the BPB correctly and you wont be able to utilize your disk to
its fullest cap.
I dont want to bore you, just wanted to give a little insight. To answer your
question, I think you should reformat your partition just to make sure.
You can off course wait till I complete my investigation on your HD. But if
there were'nt any problems reported, I strongly beleive that you should try it
again. If it failed then, that's bug and should be reported to the support
line.

I will let you know latter today about any problems with you hard disk
controller. Please let me know if I can be of any more help.

         Thanks


                   Kamran R. [IBM]

* Reply:  12965

#: 12965 S4/General 2.x Q&A
    04-Aug-92  12:05:06
Sb: #12919-#OS/2 2.0 & SCSI 1.5 Gb
Fm: Wayne Kovsky [KSW] 76164,3504
To: Kamran Rejaee [IBM] 70733,1401 (X)

Thanks, Kamran, for your informative response.

As to problems with FDISK/FDISKPM, they were reported in this thread by John
Bridges of GRASP; sorry, I can't point you to the message number.

If you find out that there is a problem with my drive/controller, please do
let me know and I will try to get it fixed, and then try to use it in "native"
mode.  Otherwise, I won't reformat my disk unless something else requires me
to.  With my "sector mapping" mode, I have a complete solution:  I can access
100% of my disk, at no penalty in performance or reliability.

In my earlier postings on this topic, I was trying to emphasize that many
people do NOT have a solution with their particular drive/controller
combinations, and they end up sacrificing a portion of their disk.  That gives
OS/2 a bad name, because these people are unhappy with that result (and the
ignore the fact that they would get exactly the same result with DOS/Windows,
and with some flavors of UNIX).  Under DOS, they get around the problem with
something like SpeedStore, but of course that doesn't run under OS/2.  I hope
you will keep pursuing this issue (perhaps by trying several different ESDI
controllers with drives that have more than 1024 cylinders), not for me - I'm
satisfied - but for other OS/2 customers or would-be customers.

* Reply:  13004

#: 13004 S4/General 2.x Q&A
    04-Aug-92  13:32:50
Sb: #12965-OS/2 2.0 & SCSI 1.5 Gb
Fm: Kamran Rejaee [IBM] 70733,1401
To: Wayne Kovsky [KSW] 76164,3504 (X)

Thanks Wayne, I'll make sure of it..