Overcome by the constant drum beat from Nico Kadel-Garcia
for moving old SCO systems to VMWare virtual machines
I followed the how-to:

http://www.blindhog.net/virtualization-vmware-player-and-easyvmx-free-virtual-machines/

To set up VMWare Player 3.1.3 build-324285 on a Windows XP
Home system running on a Gigabyte GA-MA785GM-US2H system
with AMD Athlon 5000 dual core CPU and embedded Realtek PCIe
GBE NIC.

I realize that VMWare Player for Windows is not a production
environment but I was seeking to get some experience with
running virtual machines.

I then played with VMWare Player by installing DSL Linux 0.9.1,
And Windows 98 as guest OS's to get the feel of it.

I chose old OS as that's what VMWare is touted as being
a solution for running on modern hardware.

Then I installed SCO 5.0.6 from bootable SCO media and the
installation went smoothly deferring network installation.

I have all 5.0.6 patches on a file system on my office 5.0.7
server and planned to copy them via the network once
I had the network running.

The first problem was running netconfig and it not identifying
the Intel Pro 1000 NIC configured using the easyvmx.com
configuration option for Intel 1000 NIC with the drivers
cooked into the basic 5.0.6 installation.

I solved that by copying the eeG driver to a DOS floppy,
using dd if=/dev/install of=drivers.flp bs=36b to copy the
DOS floppy to an image file that VMWare Player can open
as the floppy on the 5.0.6 guest OS (my WinXP system is sans
floppy drive).

I installed the eeG network drivers using custom and then ran
netconfg and I see the Intel Pro 1000 VMWare supplied NIC
and assigned it an IP address on my network. I had created the
5.0.6 VMX file with the NIC in "bridged" mode so that it will
use it's own IP address and not the IP of the WinXP host's NIC.

I was then able to copy patchck.tar from the 5.0.7 box to the
5.0.6 guest host but noticed that network performance seemed
slow and flaky: Telnet or rlogin between the 5.0.6 and 5.0.7
boxes experiences annoying delay in keystrokes.


I ran patchck on the 5.0.6 Guest and it appeared to run
identifying that no patches were installed and offering the
list:

INSTALLED currently on unixvm.testdom.com
--------------------------------------------------------------------
(END)

------------------------- DOWNLOAD  PACKAGES ---------------------------
MISSING packages available for DOWNLOAD:
------------------------- DOWNLOAD  PACKAGES ---------------------------
1 --- rs506a            Release Supplement 506a
2 --- oss636a           OSS636a - TCP Security supplement for 5.0.6
3 --- SYSINFO           sysinfo 11022402 - Information Gathering Tool
4 --- OSS640A           OSS640a - Bind supplement for 5.0.x
5 --- OSS642A           OSS642a - Cron supplement for 5.0.x
6 --- OSS643A           OSS643a - Socket Driver supplement for 5.0.6
7 --- OSS646C           OSS646c - Execution Env supplement for 5.0.x
8 --- OSS648C           OSS648c - Processor supplement for 5.0.6
A --- ALL               DOWNLOAD ALL of the packages above
Q --- Quit              Skip DOWNLOAD Phase
------------------------- DOWNLOAD  PHASE ---------------------------
List packages to DOWNLOAD [1,2,...,6], [A], or [Q]
     [Default=ALL]

But when I typed 1, to download rs506a it immediately went to

------------------------- DOWNLOAD PHASE ---------------------------
DOWNLOAD Phase complete..  Press <ENTER> to continue



REVIEW the logs in /var/spool/patchck/logs?
       [Y/N Default=N] y

update.log - Sat Mar 19 16:31:30 CST 2011 - patchck version 11022402
Attempting ftp connection to 132.147.127.12 for self-update
Verbose mode on.
prompt
Interactive mode off.
cd /pub/tools
250 CWD command successful.
get patchck.custom.Z
local: patchck.custom.Z remote: patchck.custom.Z
227 Entering Passive Mode (132,147,127,12,5,242)
150 Opening BINARY mode data connection for patchck.custom.Z (120113 bytes).
226 Transfer complete.
120113 bytes received in 12.73 seconds (9.21 Kbytes/s)
get patchck.version
local: patchck.version remote: patchck.version
227 Entering Passive Mode (132,147,127,12,5,244)
unable to reach 132.147.127.12 within 20 seconds
Attempting ftp connection to 132.147.127.12 for self-update
Verbose mode on.
prompt
Interactive mode off.
cd /pub/tools
250 CWD command successful.
get patchck.custom.Z
local: patchck.custom.Z remote: patchck.custom.Z
227 Entering Passive Mode (132,147,127,12,4,173)
150 Opening BINARY mode data connection for patchck.custom.Z (120113 bytes).
unable to reach 132.147.127.12 within 20 seconds
(END)

So patchck was having trouble and I would be unable to use it
to get the 5.0.6 patches.

I burned my 5.0.6 driver and patch library to cdr and mounted
the cd on the 5.0.6 guest machine and copied the files
to /usr/lib/drivers.

I ran netconfig and removed the eeG NIC, ran custom and removed
the eeG package then ran the custom installation for rs506a.

With rs506a installed, I re-installed the eeG driver and tried
patchck again with the same results.

My 5.0.6 driver library was missing oss637a and oss640a and
so I downloaded them to winXP box, copied them to /tmp on the
5.0.7 box and tried ftp from the 5.0.7 to 5.0.6 guest OS
and noted that the ftp transfer was very slow:

ftp> put oss640a.001.000
local: oss640a.001.000 remote: oss640a.001.000
200 PORT command successful.
150 Opening BINARY mode data connection for oss640a.001.000.
226 Transfer complete.
741376 bytes sent in 277.67 seconds (2.61 Kbytes/s)
ftp> put oss640a.000
local: oss640a.000 remote: oss640a.000
200 PORT command successful.
150 Opening BINARY mode data connection for oss640a.000.
226 Transfer complete.
1338368 bytes sent in 630.26 seconds (2.07 Kbytes/s)
ftp>

The above log for oss640a was after I had transfered and
installed oss636a.


Doing rlogin from 5.0.7 to the 5.0.6 guest named unixvm
and executing time cat /etc/ptest (My printer test file
of 768 lines) stops scrolling at line 391 with a long
delay and then finishes showing over one minute real time!

# time cat /etc/ptest
      1  1 2 3 4 5 6 7 8 910 1 2 3 4 5 6 7 8 920 1 2 3 4 5 6 7 8 930 1 2 3 4 5
      2  1 2 3 4 5 6 7 8 910 1 2 3 4 5 6 7 8 920 1 2 3 4 5 6 7 8 930 1 2 3 4 5
      3  1 2 3 4 5 6 7 8 910 1 2 3 4 5 6 7 8 920 1 2 3 4 5 6 7 8 930 1 2 3 4 5
      4  1 2 3 4 5 6 7 8 910 1 2 3 4 5 6 7 8 920 1 2 3 4 5 6 7 8 930 1 2 3 4 5
....
    764  1 2 3 4 5 6 7 8 910 1 2 3 4 5 6 7 8 920 1 2 3 4 5 6 7 8 930 1 2 3 4 5
    765  1 2 3 4 5 6 7 8 910 1 2 3 4 5 6 7 8 920 1 2 3 4 5 6 7 8 930 1 2 3 4 5
    766  1 2 3 4 5 6 7 8 910 1 2 3 4 5 6 7 8 920 1 2 3 4 5 6 7 8 930 1 2 3 4 5
    767  1 2 3 4 5 6 7 8 910 1 2 3 4 5 6 7 8 920 1 2 3 4 5 6 7 8 930 1 2 3 4 5
    768  1 2 3 4 5 6 7 8 910 1 2 3 4 5 6 7 8 920 1 2 3 4 5 6 7 8 930 1 2 3 4 5

real     1:34.74
user        0.00
sys         0.16
#


Time cat /etc/ptest when logged in via TermLite to the 5.0.7 box shows:

    767  1 2 3 4 5 6 7 8 910 1 2 3 4 5 6 7 8 920 1 2 3 4 5 6 7 8 930 1 2 3 4 5
    768  1 2 3 4 5 6 7 8 910 1 2 3 4 5 6 7 8 920 1 2 3 4 5 6 7 8 930 1 2 3 4 5

real    0m0.06s
user    0m0.00s
sys     0m0.06s
#




Time cat /etc/ptest on the 5.0.6 Guest console shows:

    767  1 2 3 4 5 6 7 8 910 1 2 3 4 5 6 7 8 920 1 2 3 4 5 6 7 8 930 1 2 3 4 5
    768  1 2 3 4 5 6 7 8 910 1 2 3 4 5 6 7 8 920 1 2 3 4 5 6 7 8 930 1 2 3 4 5

real    0m0.86s
user    0m0.00s
sys     0m0.86s
#

After installing all recommended patches and OSS657A, the Minimum
PIT Count Supplement the system still has slow network peformance
and patchck does not work.

INSTALLED currently on unixvm.testdom.com
--------------------------------------------------------------------
oss636a                 OSS636a - TCP Security supplement for 5.0.6
oss640a                 OSS640a - Bind supplement for 5.0.x
oss642a                 OSS642a - Cron supplement for 5.0.x
oss643a                 OSS643a - Socket Driver supplement for 5.0.6
oss646c                 OSS646c - Execution Env supplement for 5.0.x
rs506a                  Release Supplement 506a

------------------------- DOWNLOAD  PACKAGES ---------------------------
MISSING packages available for DOWNLOAD:
------------------------- DOWNLOAD  PACKAGES ---------------------------
1 --- SYSINFO           sysinfo 11022402 - Information Gathering Tool
2 --- OSS648C           OSS648c - Processor supplement for 5.0.6
A --- ALL               DOWNLOAD ALL of the packages above
Q --- Quit              Skip DOWNLOAD Phase
------------------------- DOWNLOAD  PHASE ---------------------------
List packages to DOWNLOAD [1],[2], [A], or [Q]
     [Default=ALL]: 1

------------------------- DOWNLOAD PHASE ---------------------------
DOWNLOAD Phase complete..  Press <ENTER> to continue



REVIEW the logs in /var/spool/patchck/logs?
       [Y/N Default=N] n

Accepting that VMWare Player on Windows XP home is not a production
system, the 5.0.6 guest install is totally unusable via network
login.

I'm open to suggestions on how to debug this installation and get
normal network performance.


-- 
                                      Steve Fabac
                                       S.M. Fabac & Associates
                                        816/765-1670

On 21/03/2011 06:13, Steve M. Fabac, Jr. wrote:
> Overcome by the constant drum beat from Nico Kadel-Garcia
> for moving old SCO systems to VMWare virtual machines

[...]

> The first problem was running netconfig and it not identifying
> the Intel Pro 1000 NIC configured using the easyvmx.com
> configuration option for Intel 1000 NIC with the drivers
> cooked into the basic 5.0.6 installation.

[...]

> I'm open to suggestions on how to debug this installation and get
> normal network performance.

When using VMware Server, I chose VMware's "flexible" network adapter 
which OSR506 recognised as the AMD PCnet adapter.

Have you seen http://communities.vmware.com/thread/192983 ?

-- 
RGB

On Mar 22, 12:34=A0pm, RedGrittyBrick <RedGrittyBr...@spamweary.invalid>
wrote:
> On 21/03/2011 06:13, Steve M. Fabac, Jr. wrote:
>
> > Overcome by the constant drum beat from Nico Kadel-Garcia
> > for moving old SCO systems to VMWare virtual machines
>
> [...]
>
> > The first problem was running netconfig and it not identifying
> > the Intel Pro 1000 NIC configured using the easyvmx.com
> > configuration option for Intel 1000 NIC with the drivers
> > cooked into the basic 5.0.6 installation.
>
> [...]
>
> > I'm open to suggestions on how to debug this installation and get
> > normal network performance.
>
> When using VMware Server, I chose VMware's "flexible" network adapter
> which OSR506 recognised as the AMD PCnet adapter.
>
> Have you seenhttp://communities.vmware.com/thread/192983?

In general, do not try to optimize hardware configurations inside of a
virtualized system. Let the virtualization server do that. And given
that OpenServer 5.0.6 is from 2000, do *NOT* expect it to perform at
full contemporary speeds, even in virtualization on a hot server.

RedGrittyBrick wrote:
> On 21/03/2011 06:13, Steve M. Fabac, Jr. wrote:
>> Overcome by the constant drum beat from Nico Kadel-Garcia
>> for moving old SCO systems to VMWare virtual machines
> 
> [...]
> 
>> The first problem was running netconfig and it not identifying
>> the Intel Pro 1000 NIC configured using the easyvmx.com
>> configuration option for Intel 1000 NIC with the drivers
>> cooked into the basic 5.0.6 installation.
> 
> [...]
> 
>> I'm open to suggestions on how to debug this installation and get
>> normal network performance.
> 
> When using VMware Server, I chose VMware's "flexible" network adapter 
> which OSR506 recognised as the AMD PCnet adapter.
> 
> Have you seen http://communities.vmware.com/thread/192983 ?
> 

I had not seen thread 192983 until you pointed it out. The
posters discuss network performance with E1000 and Openserver 6.0.

To continue documenting my experimenting with VMWare Player,
I created a 5.0.7 vm using the options available with the
VMWare player "create a new virtual machine" selection and clicked
through the options coming to the "customize hardware" screen.

The NIC options in the integrated VMWarePlayer New Virtual machine
wizard does not give the option of selecting the NIC emulation
and just has check boxes for "bridged," "NAT," and "Host-only."

I installed 5.0.7 from a downloaded ISO named OpenServer-5.0.7Hw-10Jun05_1800.iso
using the option to set the CD/DVD to use an ISO image file.

I started the new VM and completed the installation of 5.0.7.

I rebooted after the installation and ran the software manager
to remove the OSS656B - OpenServer Licensing Update (the kernel
was relinked) and then I rebooted the VM. 5.0.7 rebooted ok.

Next, I ran custom and installed SMP from the
OpenServer-5.0.7Hw-10Jun05_1800.iso image, changed the
VMWare Player setting to switch to an ISO containing the
VOL files for MP5 and mounted the "CD" on /mnt. I ran
custom and installed MP5. That's where the test install
of 5.0.7 fails as it will no longer boot, justs gets to
hardware splash screen and then reboots after displaying
the section for initializing the hard disk. Booting
unix.old does the same thing. Booting unix.safe gets
to the "press ctl-D or enter root password" seems to
continue after pressing ctl-D but hangs up.

The only thing that worked at this point was to boot
unix.install, enter maintenance mode, run custom
and remove SMP. Exiting custom relinked the kernel.
The system booted ok with SMP removed.

I ran netconfig and it identified the adapter:

+----------------------- Network Configuration Manager -----------------------+
|+---------------------------------------------------------------------------+|
|| Hardware   Protocol                                                  Help ||
|+---------------------------------------------------------------------------+|
|Current networking configuration:                                            |
|+---------------------------------------------------------------------------+|
||*HW SCO TCP/IP Loopback driver                                             ||
||      SCO TCP/IP                                                           ||
||      - SCO NFS Runtime System                                             ||
|| HW AMD PCNet-PCI Adapter Compatible - PCI Bus# 2,Device# 0,Function#      ||
||      SCO TCP/IP                                                           ||
||      - SCO NFS Runtime System                                             ||
||                                                                           ||

And the new kernel with the adapter booted ok:


# hwconfig -h
device      address    vec dma comment
======== ============= === === ================================================
kernel         -        -   -  rel=3.2v5.0.7 kid=2003-02-18
cpu            -        -   -  unit=1 family=15
cpuid          -        -   -  unit=1 vend=AuthenticAMD tfms=0:15:4:2
fpu            -       13   -  unit=1 type=80387-compatible
pci       0xcf8-0xcff   -   -  am=1 sc=0 buses=35
PnP            -        -   -  nodes=0
clock          -        -   -  type=TSC/2.209786106Ghz
serial    0x3f8-0x3ff   4   -  unit=0 type=Standard nports=1 base=0 16550A/16
serial    0x2f8-0x2ff   3   -  unit=1 type=Standard nports=1 base=8 16550A/16
console        -        -   -  unit=vga type=0 num=12 scoansi=1 scroll=50
floppy    0x3f2-0x3f7   6   2  unit=0 type=135ds18
kbmouse    0x60-0x64   12   -  type=Keyboard|PS/2 mouse (wheel) id=0x03
udi            -        -   -  UDI environment
adapter        -        -   -  ha=0 type=usb_msto UDI SCSI HBA
adapter   0x1f0-0x1f7  14   -  type=IDE ctlr=0 dvr=wd
adapter   0x170-0x177  15   -  type=IDE ctlr=1 dvr=wd
pnt      0x2000-0x2012 10   -  pnt0 Ver: 4.06 addr=00:0c:29:0c:ea:4f
cd-rom         -        -   -  type=IDE ctlr=1 cfg=mst unit=0 dvr=Srom->wd
disk      0x1f0-0x1f7  14   -  type=W0/0 unit=0 cyls=520 hds=128 secs=63
#

An ftp connection from my office 5.0.7 system to perform
the transfer of the oss640a files that were so slow with the
e1000 VMWare nic on 5.0.6 test VM were suitably speedy:

220 gnsvm FTP server (Version wu-2.6.2(1) Thu Aug 3 02:08:20 EDT 2006) ready.
Name (gsvm:smf): root
331 Password required for root.
Password:
230 User root logged in.
Remote system type is UNIX.
Using binary mode to transfer files.
ftp> cd /tmp
250 CWD command successful.
ftp> put oss640a.001.000
local: oss640a.001.000 remote: oss640a.001.000
200 PORT command successful.
150 Opening BINARY mode data connection for oss640a.001.000.
226 Transfer complete.
741376 bytes sent in 0.32 seconds (2227.69 Kbytes/s)
ftp> put oss640a.000
local: oss640a.000 remote: oss640a.000
200 PORT command successful.
150 Opening BINARY mode data connection for oss640a.000.
226 Transfer complete.
1338368 bytes sent in 0.18 seconds (7181.32 Kbytes/s)
ftp>


And timing cat /etc/ptest is suitably fast:

# time cat /etc/ptest

....
    767  1 2 3 4 5 6 7 8 910 1 2 3 4 5 6 7 8 920 1 2 3 4 5 6 7 8 930 1 2 3 4 5
    768  1 2 3 4 5 6 7 8 910 1 2 3 4 5 6 7 8 920 1 2 3 4 5 6 7 8 930 1 2 3 4 5

real        0.23
user        0.00
sys         0.19
#

I need to go back to the 5.0.6 test vm and see if I can switch
to the "flexible" NIC as you suggested for 5.0.6 in your
reply.


Back to the issue of SMP on VMWare Player: Googling *SCO smp under VMWare*
(no quotes) brings up a lot of traffic on the VMWare communities
forum indicating that other people have had problems with 5.0.7 (and
5.0.6) with SMP.

One comment that seems germane is the question as to whether VMWare
Player is presenting a multi core AMD Athlon 5000 CPU as two
single-core CPU's or as a  Multi core CPU.

Since I deferred licensing SMP expecting MP5 to enable multi-core
SMP. If VMWare Player is presenting two individual CPU's to
5.0.7, that may well explain why the system boots in a loop.

I'll try re-installing SMP with a available demo license to see
if that fixes the problem.


-- 
                                      Steve Fabac
                                       S.M. Fabac & Associates
                                        816/765-1670

Steve M. Fabac, Jr. wrote:
> Back to the issue of SMP on VMWare Player: Googling *SCO smp under VMWare*
> (no quotes) brings up a lot of traffic on the VMWare communities
> forum indicating that other people have had problems with 5.0.7 (and
> 5.0.6) with SMP.
> 
> One comment that seems germane is the question as to whether VMWare
> Player is presenting a multi core AMD Athlon 5000 CPU as two
> single-core CPU's or as a  Multi core CPU.


I don't have any experience with VMs, but I'd try either of these 
commands to show how many cpus the SCO OS sees:

# uname -X

# psrinfo

scot

On Mar 23, 12:39=A0pm, Scot Jenkins <sco-li...@thinkunix.net> wrote:
> Steve M. Fabac, Jr. wrote:
>
> > Back to the issue of SMP on VMWare Player: Googling *SCO smp under VMWa=
re*
> > (no quotes) brings up a lot of traffic on the VMWare communities
> > forum indicating that other people have had problems with 5.0.7 (and
> > 5.0.6) with SMP.
>
> > One comment that seems germane is the question as to whether VMWare
> > Player is presenting a multi core AMD Athlon 5000 CPU as two
> > single-core CPU's or as a =A0Multi core CPU.
>
> I don't have any experience with VMs, but I'd try either of these
> commands to show how many cpus the SCO OS sees:
>
> # uname -X
>
> # psrinfo
>
> scot

Also, try booting it with a live CD, such as Knoppix, and you can
review the configuration in a non-license sensitive fashion.