"Other" OSs that VBox can run

[mpack]

This topic is informational, it is not a support topic. Please search for previous discussions of OS/2 installation problems and create a specific discussion topic if the question has not been answered in the past.

[dlharper]

On the list of guest operating systems at https://www.virtualbox.org/wiki/Guest_OSes there are a small number which are marked with a red square and the comment "Doesn't work". Near the bottom of the list is "Visopsys", which is marked in this way. Although it is hardly a major operating system, I feel I must correct this statement.

Visopsys works perfectly well in VirtualBox. There is just one small glitch, which is easy enough to deal with once you realise what it is.

Visopsys runs without problem from floppy disk (but with a limited collection of utilities and programs) or from CD-ROM (but read-only). To use it properly you have to install it to a hard disk, and that is where the glitch shows up. If you try to install it to a "virgin" hard disk, then the resultant system will not boot.

The problem is caused by the partition routine in Visopsys which does not write the Master Boot Record of the hard disk correctly. The solution is first to create a partition on the hard disk using something else. (I tried it by booting using the Windows 98 start-up floppy, and then using FDISK from there, but I would expect other "fdisks" to do the same.) Once this has been done, then you can install Visopsys on the partition, and it will work fine.

This isn't a VirtualBox issue at all. The same happens if you try to install Visopsys to a new hard disk on a physical machine. It is simply that no-one is ever likely to do that, so the problem doesn't often arise there. When using a VM you are far more likely to start with a truly blank disk.

The system is one individual's hobby project, so it is rather limited and there is not a lot you can do with it once it is installed. But none of the issues are anything to do with virtualisation. It really ought to be listed as "works without additions", while noting the above partitioning point.

[klaus]

Thanks for the information, updated the list.

[mpack]

Well, while we're on the subject of that list, I noticed a while back that the Win98 entry has a comment which reads "Slow because VirtualBox not optimized for it. Disable VT-x.". It's the latter part that surprises me because it's the opposite of my own finding, i.e. I found that Win98 suddenly became pig slow around the time v4.0 was introduced, but it's actually very acceptable if you enable VT-x, and of course have decent VESA driver installed etc. I did wonder if this advice had been accidentally reversed.

[michaln]

Well, while we're on the subject of that list, I noticed a while back that the Win98 entry has a comment which reads "Slow because VirtualBox not optimized for it. Disable VT-x.". It's the latter part that surprises me because it's the opposite of my own finding, i.e. I found that Win98 suddenly became pig slow around the time v4.0 was introduced, but it's actually very acceptable if you enable VT-x, and of course have decent VESA driver installed etc. I did wonder if this advice had been accidentally reversed.

It's not so straightforward. With the default VGA driver, it's going to be faster without VT-x for sure. With a "proper" driver which writes to a linear framebuffer rather than to the VGA aperture, VT-x is going to be faster.

[jreagan]

No replies for 7 years, eh?

The OpenVMS operating system has been ported to x86-64 and can boot on VBox.

Code: Select all   Expand viewCollapse view

$ show cpu

System: X86VMS   Hardware type: Intel X86_64
A virtual machine running under VirtualBox
$ show system
OpenVMS XFM4-N4A  on node X86VMS    1-JUL-2020 20:19:27.25   Uptime  0 00:19:55
  Pid    Process Name    State  Pri      I/O       CPU       Page flts  Pages
00000081 SWAPPER         HIB     16        0   0 00:00:00.00         0     17
00000084 LANACP          HIB     14       64   0 00:00:00.03       443    481
00000085 FASTPATH_SERVER HIB     10        8   0 00:00:00.01       381    419
00000086 IPCACP          HIB     10        8   0 00:00:00.01       184    222
00000087 ERRFMT          HIB      7       71   0 00:00:00.06       487    526
00000089 OPCOM           HIB      8       31   0 00:00:00.07       240    228
0000008A AUDIT_SERVER    HIB     10       59   0 00:00:00.03       343    383
0000008B JOB_CONTROL     HIB      8       24   0 00:00:00.01       203    241
0000008E SMHANDLER       HIB      8       50   0 00:00:00.01       551    553
0000008F SYSTEM          CUR   0  7      170   0 00:00:00.11      1750    509
$

[mpack]

VMS? Wow, that takes me back - wow again - about 40 years! Can it do anything useful? ISTR the old company VAX did have hard disks, probably had 8 inch floppies somewhere too, but the main storage medium was big tape drives, the main input medium was punched cards. And of course it WAY predates even the most basic standards we take for granted today, like networking (especially no internet), CD-ISO, USB. And no graphics either - the VAX was a number cruncher, presentation was left to the terminals (similar to "thin clients" today but even more basic).

[jreagan]

VMS continued from the 32-bit VAX to the 64-bit Alpha and 64-bit Itanium. Now onto x86-64 (both real hardware and VMs).

While not as popular as it once was, there is still a loyal niche customer base (many of which you'd recognize but we normally don't share our customer names in public).

It is the full OS with access to network, storage, etc. As you mentioned, VMS is more of a server OS although DEC did have various VAX and Alpha workstations with a limited assortment of graphics cards. We still build the X11 server/client code. For our x86 systems, we'll just support any base graphics as a basic console. I don't see us supporting anything fancy for either GPU gaming or other GPU-based software (get your bitcoin elsewhere). Most of the later VAX systems (and all of the Alpha/Itanium systems) have CD/DVD for software installation, USB, lots of PCI/PCIx slots for NICs, storage arrays, etc. You used one of the card readers? I never saw a real one on a DEC machine (I started at DEC in 1983). I only used on on a CDC mainframe. And we still have most of our legacy compilers available (we use LLVM as a code generator): BASIC, C, COBOL, Fortran, Pascal, and our proprietary language BLISS. [They are currently just cross-compilers from the Itanium systems, but we'll have native ones within the next year.]

No VAX ran from floppies. The original VAX-11/780 had a RX01 8" floppy inside the CPU cabinet to boot the LSI-11 and to load the microcode. The OS already ran from some sort of disk (RM05, RP06, RP07 at the high end and RL02 for some of the smaller VAX systems). The first VAX shipped with a minimum of 256Kbytes (yes KILO bytes) of memory upto a max of 8MB (yes MEGA bytes) although most customers couldn't afford the dense memory boards. Even the compiler group had just a few 11/780s each with 2MB of memory each. It would support several dozen simultaneous users easily. It is nice to see it now on systems with more memory than the entire building contained in 1983. One of our current x86-64 test real hardware platforms is configured with 3TB of memory.

[mpack]

To be clear: I'm no VAX expert. I was a lowly acolyte working in the Computing Department for about 6 months in 1980. Unlike some of the other acolytes I had studied programming at Uni (using PDP-7, PDP-11 and DEC-20), so I could recognize some of the equipment around me, but there wasn't much call for programmers back then to fill the priesthood attending the great machine, this was the closest job I'd found. Mostly what I did was move punch cards from one pile to another and take completed mylar drawings out of the plotter while trying to avoid shocks of static electricity. It's not a time I remember fondly...

After six months I moved to another job within the same company: someone had gotten themselves something called an "Apple II" and needed someone with programming experience to help run it, bearing in mind you had to write all your own software back then. Heaven. I never looked back (until now!).

About 18 months into my new job the Computer Department was reforming, trying to be more responsive to internal company needs, I was asked to look at porting some of my Apple II programs (written in UCSD Pascal) back to the VAX. I played around with VMS quite a bit at that time. I remember being impressed by the command line syntax, so much smarter than anything I'd seen before. Unfortunately the VAX was quite terrible at providing the interactive experience of a personal computer, and the attempt brought it to it's knees.

I left that company for pastures new in 1983, by which time lots of people had some kind of DEC PC (I forget the name, it was probably the DEC Rainbow) on their desk, the Computing Department was in a serious decline - though soon to rise again pheonix like as "IT support". The first IBM PC appeared right about the time I was leaving. I've hardly used anything but DOS and Windows since then.

So, as to VMS... bit of a strange duck then. VMS with a GUI wouldn't be much of a nostalgia trip for me, but I guess I might take a look. BTW, does "OpenVMS" imply free to use?

[jreagan]

VMS is still mainly a DCL command line OS. We do have a Visual Studio Code plugin for IDE development.

The "Open" was mostly a stupid marketing term at the time. The code base is still closed (HPE technically owns the IP but VSI is the exclusive licensee). There is a hobbyist version that is currently transitioning from HPE-overight to VSI-oversight (vmssoftware.com). It is Alpha/Itanium at the moment. There are some Alpha emulators in the PC space that can be used to boot it. Our x86 on VBox is just entering field test and not generally available download. I expect that we will have some sort of free/hobbyist version in the next year or so but the details have yet to be worked out.