Digging out some history: trying to install SuSE 5.3 in vmware player6

English Virtualization
1

Hi,

Well, I got it running :slight_smile:

But, no network. No mouse. Sax crashes.

The most important issue is the network. The “bios” says the card is an
AMD Am79C970A, which of course, is not in the list of such an ancient
distribution (Sept 1998).

Vmware player does not allow me to choose network card type. From
another guest, running “13.1”, I get this info:


Eleanor4:~ # hwinfo --netcard
52: PCI 201.0: 0200 Ethernet controller
[Created at pci.319]
Unique ID: rBUF.5dU8kR7eh2C
Parent ID: 7EWs.3XzZP_1GzlF
SysFS ID: /devices/pci0000:00/0000:00:11.0/0000:02:01.0
SysFS BusID: 0000:02:01.0
Hardware Class: network
Model: "VMWare PRO/1000 MT Single Port Adapter"
Vendor: pci 0x8086 "Intel Corporation"
Device: pci 0x100f "82545EM Gigabit Ethernet Controller (Copper)"
SubVendor: pci 0x15ad "VMWare, Inc."
SubDevice: pci 0x0750 "PRO/1000 MT Single Port Adapter"
Revision: 0x01
Driver: "e1000"
Driver Modules: "e1000"
Device File: eth0
Memory Range: 0xfd5c0000-0xfd5dffff (rw,non-prefetchable)
Memory Range: 0xfdff0000-0xfdffffff (rw,non-prefetchable)
I/O Ports: 0x2000-0x3fff (rw)
Memory Range: 0xe7b00000-0xe7b0ffff (ro,non-prefetchable,disabled)
IRQ: 19 (26607 events)
HW Address: 00:0c:29:5f:45:90
Link detected: yes
Module Alias: "pci:v00008086d0000100Fsv000015ADsd00000750bc02sc00i00"
Driver Info #0:
Driver Status: e1000 is active
Driver Activation Cmd: "modprobe e1000"
Config Status: cfg=no, avail=yes, need=no, active=unknown
Attached to: #17 (PCI bridge)
Eleanor4:~ #

So… what can I do to get networking in that thing?

One idea is to change the virtual hardware of the virtual machine. This
is little documented. The .vmx contains this entry:

virtualHW.version = “10”

So I could start trying versions till one works… :-?

…]

Mmmm… clue… look at /var/log/boot.msg, because it reports probes on
hardware.

Wow. Got mouse! ps/2 type.

And wow! I got network!

And I can login via /telnet/ from host… X’-)
I don’t know if ssh is available, I’ll look.

Using HW version 8, and network card “AMD PCI PCnet32 (PCI bus NE2100)”.
I guess that it will even work on HW version 10 as well. I’ll try. …]
Yes, it does…

One thing about this version is that there is no “hwinfo”, so I don’t
know how to find out what it knows about the hardware.

Now I have to find out if there are still out there copies of the
updates, to update this thing - before trying to solve problems that
were solved at the time with updates.

–
Cheers / Saludos,

Carlos E. R.
(from 13.1 x86_64 “Bottle” at Telcontar)

2

On general principles (side stepping avoiding a direct answer)

  • VMware Player is an app that provides <just> enough to deploy most VMs. If you want the full blown, full featured app you will probably find Workstation (which costs a fair amount) far more suitable. Note that if you know how to configure VMware without a GUI and can ferret out all that’s required to configure, it’s perfectly possible to avoid using Workstation and use only Player. But there are few super stars who live and breath VMware to that degree.

  • My guess in any case is that you’ll likely have better success deploying in QEMU with its full virtualization (not paravirtualized) architecture. You should have a full array of options to deploy every piece of hardware used when that software was prominent, but virtualized. If you decide to do this, I doubt there should be any issue running QEMU simultaneously on a system running another virtualization technology, and unlike Xen does not require building a custom kernel.

TSU

3

On 2014-05-28 06:46, tsu2 wrote:
>
> On general principles (side stepping avoiding a direct answer)
>
> - VMware Player is an app that provides <just> enough to deploy most
> VMs.

Yes, that’s true. For instance, it does not have an option to change the
“hardware” version, but you can do it in a config file.

> If you want the full blown, full featured app you will probably
> find Workstation (which costs a fair amount) far more suitable.

Heh, I can not justify that expense for my private use.

> - My guess in any case is that you’ll likely have better success
> deploying in QEMU with its full virtualization (not paravirtualized)
> architecture. You should have a full array of options to deploy every
> piece of hardware used when that software was prominent, but
> virtualized. If you decide to do this, I doubt there should be any issue
> running QEMU simultaneously on a system running another virtualization
> technology, and unlike Xen does not require building a custom kernel.

Interesting. I will have to study that.

You know that one of the reasons I use vmware and not virtualbox, is
that it explicitly supports MsDOS and other old oses.

I got SuSE 5.3 running, by the way. The missing part is graphics, (and
there will not be guest tools unless I build them).

I don’t remember how updating was done with 5.3, but certainly not
automatic: it looks at the CD as update source, with the intention to
update to the next SuSE release. I was a novice at the time, but I think
we did them manually, from a list, using plain rpm.

I have located a source of all original packages of 5.3 at
http://ftp5.gwdg.de/pub/linux/suse/5.3/, but there is no
updates tree.

Another detail is that at one point the install wanted the CD number 3,
which does not exist in the original package I have - and I don’t find
it on gwdg.

–
Cheers / Saludos,

Carlos E. R.
(from 13.1 x86_64 “Bottle” at Telcontar)

4

:slight_smile:
5.3 was before my time. Although I dabbled in a number of distros then, I got really serious with openSUSE only after Novell acquired the technology and released its first <really> stable re-build (10.3). Everything before 10.x is the “old code,” the original OpenSuSE german code.

Yup, I’d opt for QEMU for something like this… fully virtualized, you could deploy the Linux kernel of the day on whatever CPU architecture (eg. 486 or 686), the video card in use at that time and likely the various other peripherals. This would likely have been your main option if your OS couldn’t recognize VMware’s virtualized peripherals. But of course, that could also mean you’d be subject to the expected potential vulnerabilities of running a Model T down an autobahn.

TSU

5

On 2014-05-28 16:26, tsu2 wrote:
>
> :slight_smile:
> 5.3 was before my time. Although I dabbled in a number of distros then,
> I got really serious with openSUSE only after Novell acquired the
> technology and released its first <really> stable re-build (10.3).
> Everything before 10.x is the “old code,” the original OpenSuSE german
> code.

Heh, but those old ones were SuSE, not openSUSE. Actually, 5.3 was
S.u.S.E. :slight_smile:

And very stable! The quality control then was tremendous.

> Yup, I’d opt for QEMU for something like this… fully virtualized, you
> could deploy the Linux kernel of the day on whatever CPU architecture
> (eg. 486 or 686), the video card in use at that time and likely the
> various other peripherals. This would likely have been your main option
> if your OS couldn’t recognize VMware’s virtualized peripherals.

Interesting. I must read about that. Sigh… I should read about a lot
of things, and not a lot of time free.

> But of
> course, that could also mean you’d be subject to the expected potential
> vulnerabilities of running a Model T down an autobahn.

Ah, but you don’t put a Ford T on the highway. First you close the
highway and then make a show of it :wink:

–
Cheers / Saludos,

Carlos E. R.
(from 13.1 x86_64 “Bottle” at Telcontar)