VMWare 3.0-1455; raw SCSI disk
Bob Sully
I have had to move my bootable partitions to my SCSI hard drive because
of hardware conflicts (bleeding-edge RAID controller for which there is
no Linux driver yet). I *know* that the the docs say that booting from
a raw SCSI drive is iffy...but I can't even get it to try; it always
tries to boot from the first IDE disk.
For various reasons, I need to have everything in raw partitions.
PLEASE ... can someone give me some idea as to when I can expect to be
able to boot from a raw SCSI drive? Is there work being done toward
this end? If not, I won't reasonably be able to use vmware, and this
thought does not make me happy.
Machine: Athlon XP 1600+ on Iwill XP333-R motherboard
Adaptec AHA2940U2W SCSI adapter with 36GB IBM SCSI-LVD HD
2 IDE 30GB HD's
512MB RAM
Mandrake 8.1/Linus's kernel 2.4.17 (yes, vmware-config.pl did work!)
Guest OS - Win98SE
SCSI disk 0:
Primary 1 - MS-DOS 6.22 (System Commander files here)
Primary 2 - Win98SE (FAT32)
Primary 3 - Win2K (NTFS)
Extended
Logical 5 - Linux /
Logical 6 - Linux /usr
Logical 7 - Linux /home
Logical 8 - Linux /root
Logical 9 - Linux /usr/local
Logical 10- Linux /opt
Logical 11- Linux /var
Logical 12- Linux /tmp
Logical 13- Linux swap
IDE disc 0:
Extended
Logical 5 - Windows D: (all FAT32)
Logical 6 - Windows E:
Logical 7 - Windows F:
Logical 8 - Windows G:
Logical 9 - Windows H:
Logical 10- Windows I:
IDE disc 1:
Extended
Logical 5 - Linux /backup
Logical 6 - Windows J: (backup)
--
________________________________________
Bob Sully - Simi Valley, California, USA
http://www.malibyte.net
"The weather is here - wish you were beautiful." - J. Buffett
Petr Vandrovec
> Hey all -
>
> I have had to move my bootable partitions to my SCSI hard drive because
> of hardware conflicts (bleeding-edge RAID controller for which there is
> no Linux driver yet). I *know* that the the docs say that booting from
> a raw SCSI drive is iffy...but I can't even get it to try; it always
> tries to boot from the first IDE disk.
Did you tried entering VMware BIOS, going to Boot order menu, expand
Hard disk, and move 'Add-on cards' above both IDE interfaces? I have
no better idea (yes, and I hope that your SCSI drive has at least one
bootable partition, otherwise BIOS will silently skip that device).
Petr
Bob Sully
Uh, yeah. About an hour after my first long-winded post, I tried going
into the VMware BIOS, and under "hard drive" there was an entry for
"VMware virtual SCSI HD" - which I moved above the IDE entries and - yes
- it boots, but now it *only* sees the C: partition on the SCSI disk and
the two (one IDE, one SCSI) CDROMs, as well as the network shares. All
of the Windoze data partitions (D: through I:) are on the two IDE HD's,
which are configured correctly, according to the editor - but aren't "seen".
????
Still working on this one...thanks!
-- Bob --
Petr Vandrovec wrote:
Bob Sully
IDE controller" wasn't enabled in Windoze' device manager (yellow
exclamation point). I removed the device, rebooted, and when win98
tried to reinstall it, I got a VMware dialog box stating that I didn't
have permissions on Partition 1 (extended) of the *second* IDE HD;
interestingly, it didn't complain about the first one. I put a few lines
in /etc/rc.d/rc.local (chmod 666) allowing read-write permissions on
both drives.
Thanks, Petr.
-- Bob --
Bob Sully
primary IDE controller again (it does this repeatedly; the controller is
"installed" but has the yellow exclamation point in Device Manager). I
have to remove, then reinstall it if I want to see my IDE drives; when I
do, VMware again tells me that there are permission problems on
"partition 1" of the second IDE drive (hdb). After the re-install, all
of the drives are visible, but if I reboot, it's back to square one.
Mandrake 8.1 has a "different" structure in the /dev directory for
devices. Permissions on the second drive are as follows:
[rcs@titanic: ~]$ lsl /dev/ide/host0/bus0/target1/lun0
total 0
drwxr-xr-x 1 root root 0 Dec 29 20:24 ../
drwxr-xr-x 1 root root 0 Dec 30 22:44 ./
brw-rw-rw- 1 root root 3, 64 Dec 31 1969 disc
brwxrwxrwx 1 root root 3, 65 Dec 31 1969 part1
brw-rw-rw- 1 root root 3, 69 Dec 31 1969 part5
brw-rw-rw- 1 root root 3, 70 Dec 31 1969 part6
brw-rw-rw- 1 root root 3, 71 Dec 31 1969 part7
brw-rw-rw- 1 root root 3, 72 Dec 31 1969 part8
brw-rw-rw- 1 root root 3, 73 Dec 31 1969 part9
brw-rw-rw- 1 root root 3, 74 Dec 31 1969 part10
brw-rw-rw- 1 root root 3, 75 Dec 31 1969 part11
My feeling is that this has to be due to some problem between VMware 3.0
and Mandrake 8.1. Anyone have a fix for this??
Thanks -- Bob
Bob Sully wrote:
> Fixed!! Found out that it wasn't seeing my IDE HD's because the "Primary
> IDE controller" wasn't enabled in Windoze' device manager (yellow
> exclamation point). I removed the device, rebooted, and when win98
> tried to reinstall it, I got a VMware dialog box stating that I didn't
> have permissions on Partition 1 (extended) of the *second* IDE HD;
> interestingly, it didn't complain about the first one. I put a few lines
> in /etc/rc.d/rc.local (chmod 666) allowing read-write permissions on
> both drives.
>
> Thanks, Petr.
--
Petr Vandrovec
>
> Actually, I thought I had this fixed, but upon restart, it lost the
> primary IDE controller again (it does this repeatedly; the controller is
> "installed" but has the yellow exclamation point in Device Manager). I
> have to remove, then reinstall it if I want to see my IDE drives; when I
> do, VMware again tells me that there are permission problems on
> "partition 1" of the second IDE drive (hdb). After the re-install, all
> of the drives are visible, but if I reboot, it's back to square one.
Can you make sure that you are not using DMA for these disks? There can be
problem with UDMA on some hosts - disable DMA on all devices in VMWare BIOS,
and also uncheck 'using DMA' on all devices in Win98.
Petr
Bob Sully
Curiouser and curiouser...
Disabling DMA didn't work. It appears now that the problem may be in
hardware recognition. Currently, I can get the Win98SE guest OS to boot
off the raw SCSI HD if I disable the IDE drives in the configuration; it
comes up, sees the IDE controller, installs it - but only the parent
device and the *secondary* controller - it sees the primary when it is
installing devices, but doesn't *actually* install it (I only see "Intel
... Bus Master IDE Controller" and "Secondary IDE Controller" in Device
Manager - no "Primary IDE Controller" at all). If I then re-install the
IDE drives in Configuration Editor, I get the error box:
Virtual Machine Kernel Stack Fault (hardware reset) .. etc.
Win98SE boots up fine and recognizes all drives when booted natively.
????
Thanks ... Bob
Petr Vandrovec wrote:
Chuck Gladu
wrote:
>Retr:
>
>Curiouser and curiouser...
>
>Disabling DMA didn't work. It appears now that the problem may be in
>hardware recognition. Currently, I can get the Win98SE guest OS to boot
>off the raw SCSI HD if I disable the IDE drives in the configuration; it
>comes up, sees the IDE controller, installs it - but only the parent
>device and the *secondary* controller - it sees the primary when it is
>installing devices, but doesn't *actually* install it (I only see "Intel
>... Bus Master IDE Controller"
I forgot, was this a dual boot system?
Because the guest does NOT have an "Intel Bus Master IDE Controller"
So you need to remove the Intel drivers and replace them with standard
IDE drivers. Unfortunately Intel (if I recall correctly) does NOT
provide any working method for uninstalling thier drivers completely.
>and "Secondary IDE Controller" in Device
>Manager - no "Primary IDE Controller" at all). If I then re-install the
> IDE drives in Configuration Editor, I get the error box:
>
>Virtual Machine Kernel Stack Fault (hardware reset) .. etc.
>
>Win98SE boots up fine and recognizes all drives when booted natively.
>
>????
>
>Thanks ... Bob
>
>
>Petr Vandrovec wrote:
>
>> Bob Sully wrote:
>>
>>>Actually, I thought I had this fixed, but upon restart, it lost the
>>>primary IDE controller again (it does this repeatedly; the controller is
>>>"installed" but has the yellow exclamation point in Device Manager). I
>>>have to remove, then reinstall it if I want to see my IDE drives; when I
>>>do, VMware again tells me that there are permission problems on
>>>"partition 1" of the second IDE drive (hdb). After the re-install, all
>>>of the drives are visible, but if I reboot, it's back to square one.
>>>
>>
>> Can you make sure that you are not using DMA for these disks? There can be
>> problem with UDMA on some hosts - disable DMA on all devices in VMWare BIOS,
>> and also uncheck 'using DMA' on all devices in Win98.
> Petr
----
Chuck Gladu
Do NOT reply to me by e-mail.
Please note: I do NOT work for Vmware.
I'm a customer just like you are.
Bob Sully
However, as I recall, if I remove the IDE controller and restart the
Win98SE guest, it comes up and re-installs the Intel parent device
driver automatically.
I'll try removing the Intel driver and then installing the "standard"
controller and see if it takes.
-- Bob --
Chuck Gladu wrote:
> On Sat, 05 Jan 2002 17:27:14 -0800, Bob Sully <r...@malibyte.net>
> wrote:
>>
>>Disabling DMA didn't work. It appears now that the problem may be in
>>hardware recognition. Currently, I can get the Win98SE guest OS to boot
>>off the raw SCSI HD if I disable the IDE drives in the configuration; it
>>comes up, sees the IDE controller, installs it - but only the parent
>>device and the *secondary* controller - it sees the primary when it is
>>installing devices, but doesn't *actually* install it (I only see "Intel
>>... Bus Master IDE Controller"
>
> I forgot, was this a dual boot system?
>
> Because the guest does NOT have an "Intel Bus Master IDE Controller"
>
> So you need to remove the Intel drivers and replace them with standard
> IDE drivers. Unfortunately Intel (if I recall correctly) does NOT
> provide any working method for uninstalling thier drivers completely.
Chuck Gladu
Intel's method for installing their driver makes it extremely
difficult to ever remove/replace it.
If I recall correctly Intel overwrites some of the INF files that the
OS put in place instead of having theirs go in as an OEM####.INF
This means, unfortunately, that Intel has redefined "standard" IDE to
be their driver instead of the OS version. This makes it hard to
remove because Intel has told the OS that *ALL* IDE controllers it
finds (no matter what type they are) are Intel Bus-Master Controllers
and should use the Intel drivers.
On Tue, 15 Jan 2002 22:21:38 -0800, Bob Sully <r...@malibyte.net>
wrote:
>Chuck: Thanks, will *try* to do this. Yes, it is a dual-boot machine.
>However, as I recall, if I remove the IDE controller and restart the
>Win98SE guest, it comes up and re-installs the Intel parent device
>driver automatically.
>
>I'll try removing the Intel driver and then installing the "standard"
>controller and see if it takes.
>
>-- Bob --
>
>
>Chuck Gladu wrote:
>
>> On Sat, 05 Jan 2002 17:27:14 -0800, Bob Sully <r...@malibyte.net>
>> wrote:
>>>
>>>Disabling DMA didn't work. It appears now that the problem may be in
>>>hardware recognition. Currently, I can get the Win98SE guest OS to boot
>>>off the raw SCSI HD if I disable the IDE drives in the configuration; it
>>>comes up, sees the IDE controller, installs it - but only the parent
>>>device and the *secondary* controller - it sees the primary when it is
>>>installing devices, but doesn't *actually* install it (I only see "Intel
>>>... Bus Master IDE Controller"
>>
>> I forgot, was this a dual boot system?
>>
>> Because the guest does NOT have an "Intel Bus Master IDE Controller"
>>
>> So you need to remove the Intel drivers and replace them with standard
>> IDE drivers. Unfortunately Intel (if I recall correctly) does NOT
>> provide any working method for uninstalling thier drivers completely.
----
Bob Sully
Well, this has become thoroughly discouraging. It's not letting me get
to the point where I can add the IDE controller, because it fails with a
virtual machine kernel stack fault error (hardware reset) - every time,
unless I use safe mode only (and it still doesn't see the IDE drives
there: it simply boots off the raw C: partition on the SCSI drive).
I have tried changing the IDE settings in the VMWare BIOS; the only
setting which recognized the "VMWare Virtual IDE Drives" is the "Auto"
setting. Switching it to PIO mode 4 (User) mode doesn't work either.
Changing read/write permissions on the partitions doesn't help either.
I'm about at the point of giving up here, which is a *real* drag; I only
use Windows occasionally, and the need to reboot to run one or two apps
depresses me immensely.
Recap: Athlon XP 1.47GHz, IWill XP333R motherboard; Host OS: Linux
Mandrake 8.1/custom kernel 2.4.17; guest - Win98SE/Win98Lite. Drive 0 -
IBM 36GB scsi; drive 1 - 30GB IDE, drive 2 - 30GB IDE.
This same setup worked *fine* when booting off an IDE raw disk, but for
various reasons, I needed to make the SCSI drive the boot disk.
Thanks for your help so far....Bob
Chuck Gladu wrote:
> Just a warning: This could be a *challenging* process.
>
> Intel's method for installing their driver makes it extremely
> difficult to ever remove/replace it.
>
> If I recall correctly Intel overwrites some of the INF files that the
> OS put in place instead of having theirs go in as an OEM####.INF
>
> This means, unfortunately, that Intel has redefined "standard" IDE to
> be their driver instead of the OS version. This makes it hard to
> remove because Intel has told the OS that *ALL* IDE controllers it
> finds (no matter what type they are) are Intel Bus-Master Controllers
> and should use the Intel drivers.