Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss
AlphaVM on OSX?
381 views
Skip to first unread message
Sum1
Sep 16, 2012, 11:44:56 PM9/16/12
to
Downloaded the Linux version for a play - would like to run VMS on OSX
that is more current than 7.3-2 (SimH emulating Alpha isn't available
yet…)
Has anyone got AlphaVM working on OSX and if so, could you please share
the details?
Cheers
PS Many thanks for those of you who helped with my previous dd problem.
Indeed I had imaged too much - I imaged the whole CD rather than just
the VMS partition, once corrected the iso worked fine :)
that is more current than 7.3-2 (SimH emulating Alpha isn't available
yet…)
Has anyone got AlphaVM working on OSX and if so, could you please share
the details?
Cheers
PS Many thanks for those of you who helped with my previous dd problem.
Indeed I had imaged too much - I imaged the whole CD rather than just
the VMS partition, once corrected the iso worked fine :)
John E. Malmberg
Sep 17, 2012, 1:07:46 AM9/17/12
to
On 9/16/2012 10:44 PM, Sum1 wrote:
> Downloaded the Linux version for a play - would like to run VMS on OSX
> that is more current than 7.3-2 (SimH emulating Alpha isn't available yet…)
>
> Has anyone got AlphaVM working on OSX and if so, could you please share
> the details?
I do not run emulators directly on the host OS, I run them in VMs that
> Downloaded the Linux version for a play - would like to run VMS on OSX
> that is more current than 7.3-2 (SimH emulating Alpha isn't available yet…)
>
> Has anyone got AlphaVM working on OSX and if so, could you please share
> the details?
are dedicated to supporting the emulator. Each VM is created with 4
virtual NICs. One set that are bridged to the host NIC, and one set
that are dedicated to an internal network.
I am currently using VirtualBox. I am seeing some problems with
clustering occasionally with the current versions as it appears that
sometimes the paravirtualized network drivers on the Microsoft Windows
VM occasionally stop working until I reboot that VM.
Also the cluster does not survive suspending and resuming the laptop
when moving it.
I have the emulated VMS systems NFS mounting storage via Cygwin/NFS
which is also backed up with a free DropBox.com storage, so that changes
can be made on the laptop or the desktop as long as both have internet
connectivity.
Regards,
-John
wb8...@qsl.network
Personal Opinion Only
Stephen Hoffman
Sep 17, 2012, 8:31:36 AM9/17/12
to
On 2012-09-17 03:44:56 +0000, Sum1 said:
> Downloaded the Linux version for a play - would like to run VMS on OSX
> that is more current than 7.3-2 (SimH emulating Alpha isn't available
> yet…)
V7.3 is the last OpenVMS VAX release. There is no V7.3-2 for OpenVMS
> Downloaded the Linux version for a play - would like to run VMS on OSX
> that is more current than 7.3-2 (SimH emulating Alpha isn't available
> yet…)
VAX; that's an OpenVMS Alpha release.
> Has anyone got AlphaVM working on OSX and if so, could you please share
> the details?
their "bare-metal" stuff a while back.
EmuVM was offering a pre-installed emulator as a package; a software
emulator that was sitting atop a Debian Linux distro.
This means you'll need to install a virtual machine in OS X, or use
Apple Boot Camp, then load the Linux-n-emulator distro package.
So to summarize, you're asking some VMS folks help with how to install
and troubleshoot Debian Linux and a Linux application, as a guest on OS
X.
Will it work? I don't know. Try it. Or ask EmuVM.
Have I tried this stack with OS X? No.
How to approach this? Roll in a VM on OS X (there are two third-party
products in common use), or set up Boot Camp and create a partition
(that's free with OS X), and then load the software guest into the VM
as you would any other Linux distro, and boot it.
None of which has anything particularly to do with comp.os.vms, either;
you're really asking in the wrong place for help.
Where's a better place? Either the Apple forums if you're running into
trouble with Boot Camp, or the forum for whichever VM package if you're
running into problems with the VM, or whatever EmuVM uses for
discussions if you're having trouble with the EmuVM "hardware". You
might get some help here with VMS itself, but possibly not as much as
you'd like if the "hardware" you're running VMS on proves to be flaky,
ill-documented or otherwise problematic.
Put another way, VMS might be your goal, but that goal is largely
irrelevent to the effort here, except when discussing potential bugs in
the emulator itself. And there's not much that the folks here can do
in the event there are bugs in the "hardware", for that matter.
--
Pure Personal Opinion | HoffmanLabs LLC
David Froble
Sep 17, 2012, 12:12:44 PM9/17/12
to
John,
After reading your post I downloaded Oracle's VirtualBox and some
extras. Haven't even tried to do any set-up.
I got the version for Windows. What I think I'm understanding is that
you run windows, then VirtualBox as an app on windows, and then guest OS
on VM partitions. Am I anywhere close to understanding this?
Do you have any info, hints, things to do / don't do, and such. This
will be my first attempt at anything like this. I've never tried to set
up a SimH system.
Any and all suggestions and hints will be appreciated.
Dave
David Froble Tel: 724-529-0450
Dave Froble Enterprises, Inc. E-Mail: da...@tsoft-inc.com
DFE Ultralights, Inc.
170 Grimplin Road
Vanderbilt, PA 15486
Stephen Hoffman
Sep 17, 2012, 12:26:01 PM9/17/12
to
John mail over on decuserve?
> After reading your post I downloaded Oracle's VirtualBox and some
> extras. Haven't even tried to do any set-up.
>
> I got the version for Windows. What I think I'm understanding is that
> you run windows, then VirtualBox as an app on windows, and then guest OS
> on VM partitions. Am I anywhere close to understanding this?
> Do you have any info, hints, things to do / don't do, and such. This
> will be my first attempt at anything like this. I've never tried to set
> up a SimH system.
binaries for Windows, and run it as an application. (Windows has
issues with the command line emulation of terminal sequences - it
doesn't - but you can run simh directly. No VM required.)
Paul Sture
Sep 17, 2012, 2:04:03 PM9/17/12
to
On Mon, 17 Sep 2012 13:44:56 +1000, Sum1 wrote:
> Has anyone got AlphaVM working on OSX and if so, could you please share
> the details?
AlphaVM is only available for Windows or Linux hosts. If you want to run
> Has anyone got AlphaVM working on OSX and if so, could you please share
> the details?
it on OS X, you will have to run it in a virtual machine which is running
Windows or Linux, or use Boot Camp to install one of those in a separate
partition.
--
Paul Sture
John E. Malmberg
Sep 17, 2012, 9:35:52 PM9/17/12
to
On 9/17/2012 11:12 AM, David Froble wrote:
>
> I tried this direct and it failed, so i'll post my questions here.
Watch the top level domains, I have not the money to actually purchase
>
> I tried this direct and it failed, so i'll post my questions here.
one. I do not know if the network TLD has been purchased.
An invalid top level domain keeps the spambots from pounding real mail
servers with invalid e-mail addresses, and because of that is the
preferred way of mangling an e-mail address.
If you just change the other parts, your mailserver still get hits with
the spam delivery attempts to the munged name, which it then has to reject.
> John,
>
> After reading your post I downloaded Oracle's VirtualBox and some
> extras. Haven't even tried to do any set-up.
>
> I got the version for Windows. What I think I'm understanding is that
> you run windows, then VirtualBox as an app on windows, and then guest OS
> on VM partitions. Am I anywhere close to understanding this?
> Do you have any info, hints, things to do / don't do, and such. This
> will be my first attempt at anything like this. I've never tried to set
> up a SimH system.
>
> Any and all suggestions and hints will be appreciated.
be enabled in the BIOS setup for best performance.
I have found that VirtualBox 3.1.8 is the only release that allowed VMS
clustering to work reliably. Since then, the network connections on the
Windows host have intermittently dropped out requiring a reboot.
Unfortunately VirtualBox 3.1.8 hostmode networking will crash Mac OS-X,
you have to use a later version for that to work.
The latest versions have allowed some communications. I am currently
using the latest release virtualbox and it seems to allow cluster
communications most of the time. Suspending the laptop tends to crash
the cluster because of communication timeouts. I am planing on
breakiing the cluster appart and using NFS to share the disks between
the VMS systems.
I am running FreeAXP on a Windows 7 64 bit VM with quad processors, and
SimH/VAX on an Ubuntu 64 bit VM with 1 processor, on a 4 processor
MacBook pro.
I found that unless I set up the Windows 7 VM to use all 4 processors
instead of just the two that FreeAXP required, the MacBook was CPU bound.
Windows 7/64 bit seems to need 40 GB system disk minimum. Linux seems
happy with 10 GB.
I created virtual disks that are dedicated to the emulators with a write
through cache, 10 GB for each system disk and 10 GB for each data disk.
At the time I did that, I think I had to use the cli based utility to
modify the drives, not the GUI.
On the VMs, you will want to run the paravirtualized drivers. With
these drivers, they just pass control to the host OS instead of running
the driver in emulation. (Simplified explanation)
I have 4 network adapters on each VM, two for the VM guest OS, and two
for the emulator. I have unbound all the windows drivers from the
network adapters dedicated to the emulator, and on linux set those
network adapters to be datalink only with no TCPIP address assigned.
This gives me 5 operating systems on one laptop, all with access to the
external network, and an internal network that is private to those systems.
I used FreeAXP because it allowed 2 NICs, and I was not aware of any
Linux hosted free AXP emulators at the time I set this up a few years back.
The use of VMs to host the emulators makes it easy to share and redirect
the network adapters, and also makes it easy to change the emulated
CD-ROM drive to different images.
I have set this up twice. The first time the MAC lost the format of the
hard drive.
This second time, I have a free Dropbox.com account bound to the storage
on the Windows 7 VM, and use Cygwin/NFS to serve the disk to VMS, so my
work volume on VMS is now continuously backed up to the cloud and also
replicated to my other VMS systems that are up.
Note TCPIP 5.7 NFS eco 1/2 will not mount Cygwin/NFS disks. I have been
given a ritual for making it work with ECO 3, but have not yet tested it.
Sum1
Sep 18, 2012, 2:20:23 AM9/18/12
to
Thanks for all the suggestions - I have run on Windows using VMware and
done the same on OSX with Fusion, either hosting Linux or Windows. I
was just wondering if anyone had moved further along so that a Linux
version could be run on OSX - similar to the was SimH for can run
directly on OSX.
Cheers
done the same on OSX with Fusion, either hosting Linux or Windows. I
was just wondering if anyone had moved further along so that a Linux
version could be run on OSX - similar to the was SimH for can run
directly on OSX.
Cheers
Rich Alderson
Sep 18, 2012, 3:23:44 PM9/18/12
to
"John E. Malmberg" <wb8...@qsl.network> writes:
> On 9/17/2012 11:12 AM, David Froble wrote:
>> I tried this direct and it failed, so i'll post my questions here.
> Watch the top level domains, I have not the money to actually purchase
> one. I do not know if the network TLD has been purchased.
> An invalid top level domain keeps the spambots from pounding real mail
> servers with invalid e-mail addresses, and because of that is the
> preferred way of mangling an e-mail address.
John,
> On 9/17/2012 11:12 AM, David Froble wrote:
>> I tried this direct and it failed, so i'll post my questions here.
> Watch the top level domains, I have not the money to actually purchase
> one. I do not know if the network TLD has been purchased.
> An invalid top level domain keeps the spambots from pounding real mail
> servers with invalid e-mail addresses, and because of that is the
> preferred way of mangling an e-mail address.
You've misunderstood something. It's not that any invalid TLD can be used.
There is actually a reserved TLD ".invalid" which is supposed to be used for
this purpose.
Using the standard, no one gets confused, as David was.
--
Rich Alderson ne...@alderson.users.panix.com
the russet leaves of an autumn oak/inspire once again the failed poet/
to take up his pen/and essay to place his meagre words upon the page...
David Froble
Sep 18, 2012, 6:02:22 PM9/18/12
to
Rich Alderson wrote:
> "John E. Malmberg" <wb8...@qsl.network> writes:
>
>> On 9/17/2012 11:12 AM, David Froble wrote:
>
>>> I tried this direct and it failed, so i'll post my questions here.
>
>> Watch the top level domains, I have not the money to actually purchase
>> one. I do not know if the network TLD has been purchased.
>
>> An invalid top level domain keeps the spambots from pounding real mail
>> servers with invalid e-mail addresses, and because of that is the
>> preferred way of mangling an e-mail address.
>
> John,
>
> You've misunderstood something. It's not that any invalid TLD can be used.
> There is actually a reserved TLD ".invalid" which is supposed to be used for
> this purpose.
>
> Using the standard, no one gets confused, as David was.
>
But .... David is ALWAYS confused ....
> "John E. Malmberg" <wb8...@qsl.network> writes:
>
>> On 9/17/2012 11:12 AM, David Froble wrote:
>
>>> I tried this direct and it failed, so i'll post my questions here.
>
>> Watch the top level domains, I have not the money to actually purchase
>> one. I do not know if the network TLD has been purchased.
>
>> An invalid top level domain keeps the spambots from pounding real mail
>> servers with invalid e-mail addresses, and because of that is the
>> preferred way of mangling an e-mail address.
>
> John,
>
> You've misunderstood something. It's not that any invalid TLD can be used.
> There is actually a reserved TLD ".invalid" which is supposed to be used for
> this purpose.
>
> Using the standard, no one gets confused, as David was.
>
Actually, I suspected that it might be ".net", (ouch, something just
smacked me on the back of my head), but being lazy, I just re-posted in
in c.o.v
Paul Sture
Sep 19, 2012, 11:05:41 AM9/19/12
to
In article <505812a7$0$11100$c3e...@news.astraweb.com>,
I suggest you ask the folks at AlphaVM. IIRC when they announced the
Windows version of AlphaVM here on comp.os.vms, several people asked for
a Linux version, and they produced one.
I have no idea if they have OS X expertise, or if the demand is there.
I know I would be happy to cast a vote for an OS X version :-)
--
Paul Sture
Windows version of AlphaVM here on comp.os.vms, several people asked for
a Linux version, and they produced one.
I have no idea if they have OS X expertise, or if the demand is there.
I know I would be happy to cast a vote for an OS X version :-)
--
Paul Sture
Paul Sture
Sep 19, 2012, 11:21:46 AM9/19/12
to
In article <k37i21$vqm$1...@dont-email.me>,
directly on Windows, Linux, OS X, and probably other systems too.
However, for VMS purposes SimH only supports VAX, not Alpha.
Do grab the latest stable version of SimH (3.9), as it has the ability
to throttle CPU for the VMS idle process. In earlier versions the VMS
idle process would happily consume host CPU cycles and set your fans
running at full tilt.
While John M clearly knows more than I do about setting up an Alpha
emulator and clustering with it under VirtualBox, feel free to ask me
via email about setting up VirtualBox itself.
P.S. What flavour of Windows are you using? It may (or may not) help to
know.
--
Paul Sture
David Froble <da...@tsoft-inc.com> wrote:
> I got the version for Windows. What I think I'm understanding is that
> you run windows, then VirtualBox as an app on windows, and then guest OS
> on VM partitions. Am I anywhere close to understanding this?
>
> Do you have any info, hints, things to do / don't do, and such. This
> will be my first attempt at anything like this. I've never tried to set
> up a SimH system.
>
> Any and all suggestions and hints will be appreciated.
As John M says, you do not need VirtualBox to run SimH. It runs
> I got the version for Windows. What I think I'm understanding is that
> you run windows, then VirtualBox as an app on windows, and then guest OS
> on VM partitions. Am I anywhere close to understanding this?
>
> Do you have any info, hints, things to do / don't do, and such. This
> will be my first attempt at anything like this. I've never tried to set
> up a SimH system.
>
> Any and all suggestions and hints will be appreciated.
directly on Windows, Linux, OS X, and probably other systems too.
However, for VMS purposes SimH only supports VAX, not Alpha.
Do grab the latest stable version of SimH (3.9), as it has the ability
to throttle CPU for the VMS idle process. In earlier versions the VMS
idle process would happily consume host CPU cycles and set your fans
running at full tilt.
While John M clearly knows more than I do about setting up an Alpha
emulator and clustering with it under VirtualBox, feel free to ask me
via email about setting up VirtualBox itself.
P.S. What flavour of Windows are you using? It may (or may not) help to
know.
--
Paul Sture
Paul Sture
Sep 19, 2012, 11:28:20 AM9/19/12
to
In article <F_SdnTGp0qCRTcrN...@mchsi.com>,
> If you just change the other parts, your mailserver still get hits with
> the spam delivery attempts to the munged name, which it then has to reject.
I have had surprisingly good success with my nospam@ address for many
years now. Yes it does exist, but the spambots (and many humans) assume
it leads nowhere. I have however come across several website or mailing
list sign up pages that won't accept it.
--
Paul Sture
"John E. Malmberg" <wb8...@qsl.network> wrote:
> An invalid top level domain keeps the spambots from pounding real mail
> servers with invalid e-mail addresses, and because of that is the
> preferred way of mangling an e-mail address.
Noted.
> An invalid top level domain keeps the spambots from pounding real mail
> servers with invalid e-mail addresses, and because of that is the
> preferred way of mangling an e-mail address.
> If you just change the other parts, your mailserver still get hits with
> the spam delivery attempts to the munged name, which it then has to reject.
years now. Yes it does exist, but the spambots (and many humans) assume
it leads nowhere. I have however come across several website or mailing
list sign up pages that won't accept it.
--
Paul Sture
Sum1
Sep 21, 2012, 7:39:56 AM9/21/12
to
emulator :)
And I do want to run a recent version of VMS, hence 8.3 or 8.4 which
SimH can't do :: - yet?
Cheers
Sum1
Sep 21, 2012, 7:40:21 AM9/21/12
to
Cheers
John E. Malmberg
Sep 21, 2012, 8:23:42 AM9/21/12
to
emulation. It is running native x86 code with traps for certain
privilege mode operations which are then handled by the hypervisor.
Only some privilege code is actually run in an emulated mode, and that
is mainly on older guest VMs that are not hypervisor aware, and do not
have paravirtualized drivers.
The current versions of Microsoft Windows and Linux will detect the
hypervisor and cooperate with it to improve the efficiency, especially
if you have paravirtualized drivers installed on the Guest VM.
You want to select a Linux that the hypervisor is aware of course so
that you know that it will cooperate. Typically you have to install a
package from the hypervisor vendor on the guest OS.
If you choose not to run the Alpha or VAX emulator in a VM, then you
loose several features:
1. Simple networking support.
2. Simple virtual CD-ROM support.
3. A simple private network between the VM, VM host, and the emulator.
And the ability to move a backup of the VM to a different host.
> And I do want to run a recent version of VMS, hence 8.3 or 8.4 which
> SimH can't do :: - yet?
do to.
Keith Parris
Sep 21, 2012, 12:41:17 PM9/21/12
to
On 9/21/2012 6:23 AM, John E. Malmberg wrote:
> No one has written SimH Alpha yet. I do not know how hard that would
> be do to.
I noticed in the SIMH 3.9 Beta Simulators kit there is a partial Alpha
> No one has written SimH Alpha yet. I do not know how hard that would
> be do to.
implementation:
http://simh.trailing-edge.com/sources/simhv39-0-beta.zip
Sum1
Sep 22, 2012, 1:50:54 AM9/22/12
to
On 2012-09-21 12:23:42 +0000, John E. Malmberg said:
> 1. Simple networking support.
> 2. Simple virtual CD-ROM support.
> 3. A simple private network between the VM, VM host, and the emulator.
I would be happy with just #1 :)
> 1. Simple networking support.
> 2. Simple virtual CD-ROM support.
> 3. A simple private network between the VM, VM host, and the emulator.
I thank you for your explanation of the hypervisor activity. I guess
what I meant to express is that I am concerned about the cpu/memory/IO
hit of running an Alpha-version of VMS in Linux in Fusion on OSX, as
compared to running it on OSX without Fusion. I expect that the
cpu/memory/IO hit of VAXVMS in Linux in Fusion would be more than just
just SimH on OSX but I have nothing empirical to back that up.
From reading the groups it appears that some (many?) have tried SimH on
OSX without running it in a VM - perhaps they could shed some light on
performance differences?
Cheers
Stephen Hoffman
Sep 22, 2012, 8:09:54 AM9/22/12
to
or via Emulation/Simulation atop Virtualization.
(There is one non-emulation/simulation virtualization option that I'm
aware of - HP OpenVMS I64 atop HP-IVM atop HP-VM - but that's not in
play here.)
Want simple? Get real hardware. Then you can follow the directions in
the documentation and - assuming working hardware - OpenVMS will work
as documented.
Want to do this the hard way? Via emulation, or via emulation atop
virtualization?
Want to compare emulation atop an OS atop virtualization atop an OS on
an OS X box?
Ok... Have at.
Want data? Run some tests.
In general...
simh VAX emulation has idle-loop detection, which means - with the
simulator (emulator) correctly configured and working - the simulator
(emulator) only really hogs a core when it's actually busy doing
something. Without that, simh eats a core. Pretty much all of it.
But on a multicore, that's less of an issue than it used to be. Most
any Mac from the last several years or so is multicore, too.
Alpha emulation atop an operating system atop a virtualization solution
atop an operating system? Yeah; that's likely going to consume
whatever you've allowed or configured the VM to consume for resources.
I wouldn't expect that to be particularly good at sipping cores. You
can determine the differences here with your VM client on OS X and
booting and running the configuration via Boot Camp.
I've run simh, as have others. Yes, on OS X. It works. It's not
speedy, but then the x86 processors are substantially faster than the
VAX cores of old, so the aggregate less-than-efficient performance of
OpenVMS VAX compares decently with the performance on a real VAX. And
the virtual networking support integrated in the most recent simh 3.9
bits is rather better than past incarnations.
I haven't run the "bare-metal" AlphaVM-free stuff. Which means I have
no way to compare that performance.
Comparing simh to AlphaVM to whatever else? Not something I've seen
any benchmarks for - there's fodder to write an article or two on
emulator/simulator/virtualization support here, or maybe characterizing
what's available to hobbyists. Why fewer benchmarks? Commercial sites
tend to be looking for a specific architecture - VAX or Alpha - and
then characterize performance from there. And there's not a huge
incentive to run and then publish benchmarks for something that's going
to be, well, slow.
Log off your news client, stop your comparison questions, go install
some software, and find out what works, what doesn't, and what is
clearly documented, what's discussed around the net, and what's not.
Learn how this stuff works. Or doesn't. Then write up a posting or
two, and explain your findings and show your data. Or - if things go
sideways - post a question or two.
But you're not really going to learn anything (particulary useful) by
posting here. Go try it. Seriously.
John E. Malmberg
Sep 22, 2012, 2:32:26 PM9/22/12
to
On 9/22/2012 7:09 AM, Stephen Hoffman wrote:
>
> Want data? Run some tests.
>
> In general...
>
> simh VAX emulation has idle-loop detection, which means - with the
> simulator (emulator) correctly configured and working - the simulator
> (emulator) only really hogs a core when it's actually busy doing
> something. Without that, simh eats a core. Pretty much all of it. But
> on a multicore, that's less of an issue than it used to be. Most any
> Mac from the last several years or so is multicore, too.
>
> Alpha emulation atop an operating system atop a virtualization solution
> atop an operating system? Yeah; that's likely going to consume whatever
> you've allowed or configured the VM to consume for resources. I wouldn't
> expect that to be particularly good at sipping cores. You can determine
> the differences here with your VM client on OS X and booting and running
> the configuration via Boot Camp.
A hypervisor aware guest OS VM does the same idle loop detection as SimH
>
> Want data? Run some tests.
>
> In general...
>
> simh VAX emulation has idle-loop detection, which means - with the
> simulator (emulator) correctly configured and working - the simulator
> (emulator) only really hogs a core when it's actually busy doing
> something. Without that, simh eats a core. Pretty much all of it. But
> on a multicore, that's less of an issue than it used to be. Most any
> Mac from the last several years or so is multicore, too.
>
> Alpha emulation atop an operating system atop a virtualization solution
> atop an operating system? Yeah; that's likely going to consume whatever
> you've allowed or configured the VM to consume for resources. I wouldn't
> expect that to be particularly good at sipping cores. You can determine
> the differences here with your VM client on OS X and booting and running
> the configuration via Boot Camp.
does. It will also turn over other functions to the host OS and with
paravirtualized drivers, does minimal emulated I/O. Yes it uses a few
more cycles than if it was not there, but with nothing going on on an
Alpha and VAX emulator both running in their own VMs. I am seeing about
92% idle on the OS-X activity monitor, and the only thing I did special
about the guest VMs is make sure that they were configured with the
paravirtualized I/O drivers. When I change the configuration to not be
clustered, I will see if it uses less CPU.
User CPU is 2.0 to 2.3 %
System CPU is 5.5 to 6.1 %
I see the most contention for disk IO. That is probably because I
configured the VMs not to cache writes.
I am also running syngery to remotely control the keyboard and mouse of
the Macbook, which is going to use some cycles.
The biggest issue that I have seen is that under some conditions, the
network connection on the Windows 7 vm stops passing traffic until I
reboot the Windows VM. I have only seen that behavior on versions of
VirtualBox greater than 3.1.8 as I posted before.
Currently I am running VirtualBox 4.2.0_RC2 r80231.
I have not yet run the VUPOMETER tests on the emulated VMS systems, but
it might be interesting to see ratings from the various devices VMS has
been emulated on.
> Log off your news client, stop your comparison questions, go install
> some software, and find out what works, what doesn't, and what is
> clearly documented, what's discussed around the net, and what's not.
> Learn how this stuff works. Or doesn't. Then write up a posting or
> two, and explain your findings and show your data. Or - if things go
> sideways - post a question or two.
>
> But you're not really going to learn anything (particulary useful) by
> posting here. Go try it. Seriously.
David Froble
Sep 22, 2012, 4:17:11 PM9/22/12
to
using Windows 2000 Pro, and the other Windows XP. I do not have any
experience with multi-core CPUs at this time. Some questions.
Will any of the OSs, Weendoze, Linux, Unix, etc allow choices for what
can use each core?
I see references to System CPU and User CPU. Is this user configurable?
If so, what OS allow this?
My choice if I was to use a VM would be to have no underlaying OS, but,
I'm getting the impression that such does not exist. If some
underlaying OS is required, is there any that are superior for doing so?
I'd think what would be desired would be something that provided
minimal capabilities, and let the VM, and guest OSs, do as much of
what's required as possible.
What underlaying OS works well with a hypervisor and paravirtulized
drivers? Or perhaps I should be asking what guest OS can work with the
paravirtulized drivers?
Basically I'm a novice at these concepts, and looking for advice to keep
me from going down some dead end paths.
What I'm considering is putting together a new system for this
experimentation. Maybe a 4 core CPU, with plenty of memory. Not sure
what effect and bonuses you get with multiple cores, and what might be a
minimum number of cores, and what might be optimum.
As to disk storage. Are virtual devices set up for the guest OSs? Is
multiple disks or partitions advisable? Or just have one large disk
that gets virtual devices created for the guests?
So, advice on:
1) Hardware
2) Underlaying OS and configuration
3) VM software
4) Guest OSs
5) configuration, configuration, cofiguration ....
John E. Malmberg
Sep 22, 2012, 8:03:25 PM9/22/12
to
On 9/22/2012 3:17 PM, David Froble wrote:
>
> My 2 most advanced PCs are AMD San Deigo chips, 90 nm single core. One
> using Windows 2000 Pro, and the other Windows XP. I do not have any
> experience with multi-core CPUs at this time. Some questions.
According to http://en.wikipedia.org/wiki/X86_virtualization
>
> My 2 most advanced PCs are AMD San Deigo chips, 90 nm single core. One
> using Windows 2000 Pro, and the other Windows XP. I do not have any
> experience with multi-core CPUs at this time. Some questions.
You need a 64 bit AMD processor released after May 23, 2006. For Intel
you need a Pentium 4 released after November 2005.
The FreeAXP emulator requires dual processors at a minimum.
Basically I would not bother with emulation on CPUs that are more than a
few years old.
> Will any of the OSs, Weendoze, Linux, Unix, etc allow choices for what
> can use each core?
> I see references to System CPU and User CPU. Is this user configurable?
> If so, what OS allow this?
User code is non-privileged code.
Some OSes like VMS allow quotas on a user CPU for a process.
And in some cases you can set priorities on the user mode CPU.
> My choice if I was to use a VM would be to have no underlaying OS, but,
> I'm getting the impression that such does not exist
distribution as the control VM.
Others add a hypervisor to a booted operating system.
The hypervisor timeshares the guest OSes, and can be an add-on to a
booted OS.
> If some
> underlaying OS is required, is there any that are superior for doing so?
can inspect the source.
> I'd think what would be desired would be something that provided
> minimal capabilities, and let the VM, and guest OSs, do as much of
> what's required as possible.
the hypervisor or has to be simulated in user mode by the hypervisor.
So the hypervisor has to do as much of the basic IO and virtual memory
management as possible, as as simulation of that code uses most of the
resources.
For efficiency, you want the VM to do as little in privileged mode as
possible.
> What underlaying OS works well with a hypervisor and paravirtulized
> drivers? Or perhaps I should be asking what guest OS can work with the
> paravirtulized drivers?
Windows 7, 2008 have paravirtualized drivers available for them and now
> What underlaying OS works well with a hypervisor and paravirtulized
> drivers? Or perhaps I should be asking what guest OS can work with the
> paravirtulized drivers?
recognize when they are running under many of the hypervisors.
Paravirtualized drivers are also there for Windows XP.
Now the paravirualized drivers have to be matched to the hypervisor, so
you need to check to see if they are available.
Most modern Linux distributions will also recognize when they are
running in a hypervisor.
> Basically I'm a novice at these concepts, and looking for advice to keep
> me from going down some dead end paths.
VirtualBox and the guest OSes. And then look at the other Hypervisors.
VirtualBox is easy to install on Windows and Mac OS-X and provides a GUI
that is easy to use.
> What I'm considering is putting together a new system for this
> experimentation. Maybe a 4 core CPU, with plenty of memory. Not sure
> what effect and bonuses you get with multiple cores, and what might be a
> minimum number of cores, and what might be optimum.
8 core CPUs when you do your research.
I have not done the research to see if a newer 4 core processor would be
better than an older 8 core processor with the same clock speed.
> As to disk storage. Are virtual devices set up for the guest OSs? Is
> multiple disks or partitions advisable? Or just have one large disk
> that gets virtual devices created for the guests?
In general, you can use storage managers to combine and partition disks
many different ways.
For reliability, mirroring can be used.
For performance, solid state caches and separate physical disks.
For just playing, you can use one large disk.
> So, advice on:
>
> 1) Hardware
and at least 4 GB of RAM would be the minium to play with.
Faster is better, more memory is better. Nothing that you do not
already know.
If you want to play with moving a VM from one host to another with out
shutting down the guest, then you need more than simple hardware.
> 2) Underlaying OS and configuration
> 3) VM software
Pick the hypervisor(s) that you want to try.
Microsoft has Hyper-V.
Linux has Xen and KVM.
Oracle has VirtualBox that can use the KVM paravirtualized drivers.
VMware has their free player.
There are probably some others out there.
For when I do something similar, I would probably use either Xen or KVM,
depending on the guest OS support. I would probably use Scientific
Linux for as for $Dayjob related reasons I am using it.
For playing, I would also tend to open source.
As far as market share, VMware appears to have the largest commercial
market share. I do not know about the others.
When you do your research, you will find that there are some diverse
opinions, just like for operating system tuning. Some people are
against using paravirtualization for various reasons. My opinion is
paravirtualize as much as possible.
Some feel that RAM should be dedicated to the guest. Some
virtualization solutions map the guest memory to their own VM and take
over the paging. The load that you are running may be the key to the
right answer to that one.
I have been told by other experts, that if you use the ZFS file system,
you need to configure it to have a high speed write cache, such as using
solid state disks, or your VMs will have severe performance problems
every time that they do a disk write. Seems to be a common
configuration problem.
> 4) Guest OSs
Check which guest OSes have support for running in a hypervisor.
> 5) configuration, configuration, cofiguration ....
have to reinstall everything.
Try out assigning different number of CPUs to guests to see what happens.
brad
Sep 22, 2012, 9:12:46 PM9/22/12
to
On 2012-09-23, John E. Malmberg <wb8...@qsl.network> wrote:
> On 9/22/2012 3:17 PM, David Froble wrote:
>>
>> My 2 most advanced PCs are AMD San Deigo chips, 90 nm single core. One
>> using Windows 2000 Pro, and the other Windows XP. I do not have any
>> experience with multi-core CPUs at this time. Some questions.
>
> According to http://en.wikipedia.org/wiki/X86_virtualization
>
> You need a 64 bit AMD processor released after May 23, 2006. For Intel
> you need a Pentium 4 released after November 2005.
Thanks for this information - unfortunately, the article goes on to explain
> On 9/22/2012 3:17 PM, David Froble wrote:
>>
>> My 2 most advanced PCs are AMD San Deigo chips, 90 nm single core. One
>> using Windows 2000 Pro, and the other Windows XP. I do not have any
>> experience with multi-core CPUs at this time. Some questions.
>
> According to http://en.wikipedia.org/wiki/X86_virtualization
>
> You need a 64 bit AMD processor released after May 23, 2006. For Intel
> you need a Pentium 4 released after November 2005.
that not all Intel processors manufactured after this date support
hardware virtualization. A more complete list (referenced in the Wiki)
can be found here:
<http://ark.intel.com/Products/VirtualizationTechnology>
The laptop I am currently using does not support the technology. : - (
Again, thanks for the explanations - my earlier questions regarding your
experiences have been adequately answered.
[...]
John Wallace
Sep 23, 2012, 8:06:09 AM9/23/12
to
On Sep 23, 1:03 am, "John E. Malmberg" <wb8...@qsl.network> wrote:
> On 9/22/2012 3:17 PM, David Froble wrote:
>
>
>
> > My 2 most advanced PCs are AMD San Deigo chips, 90 nm single core. One
> > using Windows 2000 Pro, and the other Windows XP. I do not have any
> > experience with multi-core CPUs at this time. Some questions.
>
> According tohttp://en.wikipedia.org/wiki/X86_virtualization
> On 9/22/2012 3:17 PM, David Froble wrote:
>
>
>
> > My 2 most advanced PCs are AMD San Deigo chips, 90 nm single core. One
> > using Windows 2000 Pro, and the other Windows XP. I do not have any
> > experience with multi-core CPUs at this time. Some questions.
>
satisfactory results, at least for testing rather than production e.g.
running SuSe 11 under VMware Player (4?) on Windows XP on home laptop
(HP 6715, AMD Turion x2, 2GB) and work desktop (Dell GX260, <2GB, at
one stage). Depends on workload, amongst other things. For an initial
play, use what's available?
Wrt VMware Player: is it still zero cost for commercial use? I'm
reasonably sure it used to be (which is why I had it both at work and
at home), but with Player 5 there seems to be a commercial licence for
commercial use?
E.g. "VMware Player is free for personal non-commercial use. VMware
Player is available for commercial use as part of VMware Fusion
Professional . VMware Fusion Professional provides a commercial
license that can be used with VMware Fusion on the Mac or with VMware
Player on Windows or Linux." from
http://www.vmware.com/products/player/faqs.html
John Wallace
Sep 23, 2012, 8:19:59 AM9/23/12
to
white.
At least one of the Alpha emulator people claims to be a "bare metal"
implementation but is actually based on a Linux subset. I forget
which.
At least one VMware product (ESX? ESXi?) claims to be a "bare metal"
implementation but has historically had Linux dependencies (which have
allegedly been removed - to be replaced by what, I'm not sure).
Either way, as the complexities of the HYPErvisors increase (resource
management, scheduling, security, mobility, networking, UIs for the
above, etc) the HYPErvisors become increasingly like a classical
operating system, with the main difference being that they generally
don't encourage you to think of it as an OS in itself.
If these things do what you need, fair enough. But don't believe the
HYPE. Not all of it anyway.
Paul Sture
Sep 24, 2012, 6:09:32 AM9/24/12
to
In article
<2a30051f-c182-4c20...@ib4g2000vbb.googlegroups.com>,
A couple of cases here.
1. A 32 bit Pentium 4 with twin 1.5 GHz CPUs and 1.5GB memory running
either XP or Linux could run another version of Linux under
VirtualBox, albeit definitely slower than on a more up to
date system. Fine for finding your way around the setup of
such a configuration though.
Running VMS 7.3 in SimH with the idle loop feature enabled is
very satisfactory. A VMS shadow copy across physical disks is
far swifter than any real VAX or Alpha running SCSI disks that
I have come across.
2. I have an HP MicroServer, which is a 2 core 1.4GHz AMD box with
4GB RAM. I have used this running Windows Home Server 2011
and more recently Fedora to host VirtualBox Windows and Linux
clients. No it isn't as fast as the 2.9 GHz 2 core AMD system
I use in a similar setup, but I suspect it has better I/O
capabilities. Both boxes run AlphaVM pretty well,
but putting Windows inside VirtualBox and then one of the
Windows Alpha emulators inside that to get the networking
capability John Malmberg describes resulted in much slower
performance.
For both those boxes, more RAM would help, but extra RAM for
that > 2 year old 2.9 GHz system was USD 50 to add 2GB to
the original 2GB, then USD 180 for one 4GB module, so we
stopped there (and looking recently hoping for a price drop,
it seems these are hard to come by).
> Wrt VMware Player: is it still zero cost for commercial use? I'm
> reasonably sure it used to be (which is why I had it both at work and
> at home), but with Player 5 there seems to be a commercial licence for
> commercial use?
>
> E.g. "VMware Player is free for personal non-commercial use. VMware
> Player is available for commercial use as part of VMware Fusion
> Professional . VMware Fusion Professional provides a commercial
> license that can be used with VMware Fusion on the Mac or with VMware
> Player on Windows or Linux." from
> http://www.vmware.com/products/player/faqs.html
I recently read the EULA for the VirtualBox extensions package and
became concerned about my own use of it. However I did some more
digging last week. With Version 4 all the closed source elements were
moved into the Extensions Pack, so as long as you stick to the Personal
or Educational Uses of that you are OK. Most folks will only need the
Extension Pack for USB 2.0 raw access or Remote Desktop Protocol (RDP)
access, and the latter is the one the EULA seems to be concerned about.
I have done a write up of that, and included a copy of the current EULA,
because I could only find an obsolete version on the VirtualBox site:
http://www.sture.ch/node/221
My interpretation of the remote access restrictions for Personal Use
is that you can stick a server in a corner and access as many VirtualBox
clients running on that as you wish from *one workstation at a time".
You can of course run VirtualBox without the Extension Pack installed.
I only installed it in the first place because I wanted to access raw
USB 2.0 devices (very useful when you are migrating large wads of data),
and to date haven't used the RDP functionality at all.
--
Paul Sture
<2a30051f-c182-4c20...@ib4g2000vbb.googlegroups.com>,
1. A 32 bit Pentium 4 with twin 1.5 GHz CPUs and 1.5GB memory running
either XP or Linux could run another version of Linux under
VirtualBox, albeit definitely slower than on a more up to
date system. Fine for finding your way around the setup of
such a configuration though.
Running VMS 7.3 in SimH with the idle loop feature enabled is
very satisfactory. A VMS shadow copy across physical disks is
far swifter than any real VAX or Alpha running SCSI disks that
I have come across.
2. I have an HP MicroServer, which is a 2 core 1.4GHz AMD box with
4GB RAM. I have used this running Windows Home Server 2011
and more recently Fedora to host VirtualBox Windows and Linux
clients. No it isn't as fast as the 2.9 GHz 2 core AMD system
I use in a similar setup, but I suspect it has better I/O
capabilities. Both boxes run AlphaVM pretty well,
but putting Windows inside VirtualBox and then one of the
Windows Alpha emulators inside that to get the networking
capability John Malmberg describes resulted in much slower
performance.
For both those boxes, more RAM would help, but extra RAM for
that > 2 year old 2.9 GHz system was USD 50 to add 2GB to
the original 2GB, then USD 180 for one 4GB module, so we
stopped there (and looking recently hoping for a price drop,
it seems these are hard to come by).
> Wrt VMware Player: is it still zero cost for commercial use? I'm
> reasonably sure it used to be (which is why I had it both at work and
> at home), but with Player 5 there seems to be a commercial licence for
> commercial use?
>
> E.g. "VMware Player is free for personal non-commercial use. VMware
> Player is available for commercial use as part of VMware Fusion
> Professional . VMware Fusion Professional provides a commercial
> license that can be used with VMware Fusion on the Mac or with VMware
> Player on Windows or Linux." from
> http://www.vmware.com/products/player/faqs.html
became concerned about my own use of it. However I did some more
digging last week. With Version 4 all the closed source elements were
moved into the Extensions Pack, so as long as you stick to the Personal
or Educational Uses of that you are OK. Most folks will only need the
Extension Pack for USB 2.0 raw access or Remote Desktop Protocol (RDP)
access, and the latter is the one the EULA seems to be concerned about.
I have done a write up of that, and included a copy of the current EULA,
because I could only find an obsolete version on the VirtualBox site:
http://www.sture.ch/node/221
My interpretation of the remote access restrictions for Personal Use
is that you can stick a server in a corner and access as many VirtualBox
clients running on that as you wish from *one workstation at a time".
You can of course run VirtualBox without the Extension Pack installed.
I only installed it in the first place because I wanted to access raw
USB 2.0 devices (very useful when you are migrating large wads of data),
and to date haven't used the RDP functionality at all.
--
Paul Sture
Paul Sture
Sep 24, 2012, 6:51:54 AM9/24/12
to
In article <4LSdnRyVIdzTmcPN...@mchsi.com>,
"configured with the paravirtualized I/O drivers"?
I have VirtualBox Guest Additions installed on clients. Is that enough,
or are there other drivers available?
> User CPU is 2.0 to 2.3 %
> System CPU is 5.5 to 6.1 %
>
> I see the most contention for disk IO. That is probably because I
> configured the VMs not to cache writes.
Even with the write cache enabled, the surest way to grind your system
to a halt is to have 2 VMs competing for disk I/O. Don't try installing
a Windows VM client at the same time as another one (or the host) is
doing backups, for example.
If you can spread I/O across disks, you will see potentially massive
gains here, and possibly turn an unworkable system into one which is
pleasant to work on. I keep my VM container files off the system disk
where possible, but the more disks the merrier. From conversations with
folks who develop against databases such as Microsoft SQL Server or
Oracle, and use VMs to accomplish that, they look for laptops which you
can fit 2 disks into (and at least one of them preferably an SSD).
> I am also running syngery to remotely control the keyboard and mouse of
> the Macbook, which is going to use some cycles.
http://synergy-foss.org
(site best viewed with Javascript turned off, to avoid invitations to
download some other product completely. which will allegedly "Speed up
your Mac")
"Synergy lets you easily share your mouse and keyboard between multiple
computers on your desk, and it's Free and Open Source. Just move your
mouse off the edge of one computer's screen on to another. You can even
share all of your clipboards. All you need is a network connection.
Synergy is cross-platform (works on Windows, Mac OS X and Linux)."
That's worth a look. Thanks for the tip.
> The biggest issue that I have seen is that under some conditions, the
> network connection on the Windows 7 vm stops passing traffic until I
> reboot the Windows VM. I have only seen that behavior on versions of
> VirtualBox greater than 3.1.8 as I posted before.
>
> Currently I am running VirtualBox 4.2.0_RC2 r80231.
>
> I have not yet run the VUPOMETER tests on the emulated VMS systems, but
> it might be interesting to see ratings from the various devices VMS has
> been emulated on.
Does anyone have a common version of VUPOMETER they agree on? Someone
mentioned the Charon site earlier this year, but I couldn't find
anything there.
Hoff wrote:
> > Log off your news client, stop your comparison questions, go install
> > some software, and find out what works, what doesn't, and what is
> > clearly documented, what's discussed around the net, and what's not.
> > Learn how this stuff works. Or doesn't. Then write up a posting or
> > two, and explain your findings and show your data. Or - if things go
> > sideways - post a question or two.
> >
> > But you're not really going to learn anything (particulary useful) by
> > posting here. Go try it. Seriously.
>
> Hands on experience is the best.
Agreed.
--
Paul Sture
"John E. Malmberg" <wb8...@qsl.network> wrote:
> On 9/22/2012 7:09 AM, Stephen Hoffman wrote:
> >
> > Want data? Run some tests.
> >
> > Want data? Run some tests.
> A hypervisor aware guest OS VM does the same idle loop detection as SimH
> does. It will also turn over other functions to the host OS and with
> paravirtualized drivers, does minimal emulated I/O. Yes it uses a few
> more cycles than if it was not there, but with nothing going on on an
> Alpha and VAX emulator both running in their own VMs. I am seeing about
> 92% idle on the OS-X activity monitor, and the only thing I did special
> about the guest VMs is make sure that they were configured with the
> paravirtualized I/O drivers. When I change the configuration to not be
> clustered, I will see if it uses less CPU.
In the VirtualBox context, how do you ensure that the guest VMs are
> does. It will also turn over other functions to the host OS and with
> paravirtualized drivers, does minimal emulated I/O. Yes it uses a few
> more cycles than if it was not there, but with nothing going on on an
> Alpha and VAX emulator both running in their own VMs. I am seeing about
> 92% idle on the OS-X activity monitor, and the only thing I did special
> about the guest VMs is make sure that they were configured with the
> paravirtualized I/O drivers. When I change the configuration to not be
> clustered, I will see if it uses less CPU.
"configured with the paravirtualized I/O drivers"?
I have VirtualBox Guest Additions installed on clients. Is that enough,
or are there other drivers available?
> User CPU is 2.0 to 2.3 %
> System CPU is 5.5 to 6.1 %
>
> I see the most contention for disk IO. That is probably because I
> configured the VMs not to cache writes.
to a halt is to have 2 VMs competing for disk I/O. Don't try installing
a Windows VM client at the same time as another one (or the host) is
doing backups, for example.
If you can spread I/O across disks, you will see potentially massive
gains here, and possibly turn an unworkable system into one which is
pleasant to work on. I keep my VM container files off the system disk
where possible, but the more disks the merrier. From conversations with
folks who develop against databases such as Microsoft SQL Server or
Oracle, and use VMs to accomplish that, they look for laptops which you
can fit 2 disks into (and at least one of them preferably an SSD).
> I am also running syngery to remotely control the keyboard and mouse of
> the Macbook, which is going to use some cycles.
(site best viewed with Javascript turned off, to avoid invitations to
download some other product completely. which will allegedly "Speed up
your Mac")
"Synergy lets you easily share your mouse and keyboard between multiple
computers on your desk, and it's Free and Open Source. Just move your
mouse off the edge of one computer's screen on to another. You can even
share all of your clipboards. All you need is a network connection.
Synergy is cross-platform (works on Windows, Mac OS X and Linux)."
That's worth a look. Thanks for the tip.
> The biggest issue that I have seen is that under some conditions, the
> network connection on the Windows 7 vm stops passing traffic until I
> reboot the Windows VM. I have only seen that behavior on versions of
> VirtualBox greater than 3.1.8 as I posted before.
>
> Currently I am running VirtualBox 4.2.0_RC2 r80231.
>
> I have not yet run the VUPOMETER tests on the emulated VMS systems, but
> it might be interesting to see ratings from the various devices VMS has
> been emulated on.
mentioned the Charon site earlier this year, but I couldn't find
anything there.
Hoff wrote:
> > Log off your news client, stop your comparison questions, go install
> > some software, and find out what works, what doesn't, and what is
> > clearly documented, what's discussed around the net, and what's not.
> > Learn how this stuff works. Or doesn't. Then write up a posting or
> > two, and explain your findings and show your data. Or - if things go
> > sideways - post a question or two.
> >
> > But you're not really going to learn anything (particulary useful) by
> > posting here. Go try it. Seriously.
>
> Hands on experience is the best.
--
Paul Sture
Paul Sture
Sep 24, 2012, 8:00:50 AM9/24/12
to
In article <k3l685$uq4$1...@dont-email.me>,
container files. Mess something up? About to do something which could
mess things up? Simple - take copies of the containers and restore as
and when needed.
I started doing this a couple of years ago, and back then my best
experience was with Windows Server 2008 as a host. But that was an
evaluation copy valid for 180 days and the real product was way too
expensive.
One really useful aspect of it is that the built in backup utility has
the ability to only backup changed blocks, so after you have run a
virtual machine which sits inside a multi-GB container file, an
incremental backup is very swift.
A cheaper alternative presented itself with Windows Home Server 2011,
which you could (still can?) get for around USD 50. That also offers
the ability to backup just blocks that have changed, and despite various
other grumbles I had with the product itself, it worked well as a VM
host. Unfortunately with Server 2012, there is no direct replacement,
so you are looking at the 500 buck mark for a basic system.
Windows 7 got much better as the patches arrived, but it still doesn't
have a decent backup utility; you'll have to roll your own or go third
party for that.
Although I have been using a variety of Linux hosts for several months
now, I haven't come to any firm conclusions yet about which is "the
best" for this kind of work. Maybe others will chip in.
Given the learning curve for Linux, it might be better for yourself to
start with Windows 7* as the host. Windows 8 is still an unknown
quantity... Another alternative while you see how things pan out would
be to get an evaluation copy of Windows Server 2008 R2, which is valid
for 180 days. You do fall into the category of evaluating :-)
If you look around, you can get boxes which come without any OS
included, and yes, from HP too.
* I learned the hard way. My first VM capable PC came with Windows Home
Premium, and I quickly learned that it contained too many restrictions
for what I wanted to do. I very reluctantly forked out for an OEM copy
of Windows 7 Ultimate, and felt I had been conned.
>
> Basically I'm a novice at these concepts, and looking for advice to keep
> me from going down some dead end paths.
>
> What I'm considering is putting together a new system for this
> experimentation. Maybe a 4 core CPU, with plenty of memory. Not sure
> what effect and bonuses you get with multiple cores, and what might be a
> minimum number of cores, and what might be optimum.
Looks good. Go for plenty of RAM too. At the moment I'd suggest 8GB as
a minimum, but look for a system which can be expanded to more, and the
more the merrier. Also look at disk and interface expansion capability.
> As to disk storage. Are virtual devices set up for the guest OSs? Is
> multiple disks or partitions advisable? Or just have one large disk
> that gets virtual devices created for the guests?
You already know the answer to this :-)
As I said in my earlier reply to John M., multiple disks are the way to
go when dealing with virtual machines. At one point I had 3 disks in a
system - the system disk, the disk for VM clients, and another to hold
installation DVD images and backups. I also have a few external USB
disks for backups. Apart from installing the host OS, you will find you
don't really need a stack of DVDs any more - you download the boot
images and point your VM software at one of those when creating a client.
Also note that once you start creating VM clients, you will eat up disk
space like you would never have believed. Now that 2TB disks are common
and 3TB disks are around, you will probably find that you can get good
prices on 1.5TB or lesser capacity disks. Just what you already know
here.
And then there's the question of using an SSD for your system disk.
These are wickedly fast in comparison with traditional disks. :-)
--
Paul Sture
David Froble <da...@tsoft-inc.com> wrote:
> My choice if I was to use a VM would be to have no underlaying OS, but,
> I'm getting the impression that such does not exist. If some
> underlaying OS is required, is there any that are superior for doing so?
One Big Plus with a host OS is that you can back up the VM client
> My choice if I was to use a VM would be to have no underlaying OS, but,
> I'm getting the impression that such does not exist. If some
> underlaying OS is required, is there any that are superior for doing so?
container files. Mess something up? About to do something which could
mess things up? Simple - take copies of the containers and restore as
and when needed.
I started doing this a couple of years ago, and back then my best
experience was with Windows Server 2008 as a host. But that was an
evaluation copy valid for 180 days and the real product was way too
expensive.
One really useful aspect of it is that the built in backup utility has
the ability to only backup changed blocks, so after you have run a
virtual machine which sits inside a multi-GB container file, an
incremental backup is very swift.
A cheaper alternative presented itself with Windows Home Server 2011,
which you could (still can?) get for around USD 50. That also offers
the ability to backup just blocks that have changed, and despite various
other grumbles I had with the product itself, it worked well as a VM
host. Unfortunately with Server 2012, there is no direct replacement,
so you are looking at the 500 buck mark for a basic system.
Windows 7 got much better as the patches arrived, but it still doesn't
have a decent backup utility; you'll have to roll your own or go third
party for that.
Although I have been using a variety of Linux hosts for several months
now, I haven't come to any firm conclusions yet about which is "the
best" for this kind of work. Maybe others will chip in.
Given the learning curve for Linux, it might be better for yourself to
start with Windows 7* as the host. Windows 8 is still an unknown
quantity... Another alternative while you see how things pan out would
be to get an evaluation copy of Windows Server 2008 R2, which is valid
for 180 days. You do fall into the category of evaluating :-)
If you look around, you can get boxes which come without any OS
included, and yes, from HP too.
* I learned the hard way. My first VM capable PC came with Windows Home
Premium, and I quickly learned that it contained too many restrictions
for what I wanted to do. I very reluctantly forked out for an OEM copy
of Windows 7 Ultimate, and felt I had been conned.
>
> Basically I'm a novice at these concepts, and looking for advice to keep
> me from going down some dead end paths.
>
> What I'm considering is putting together a new system for this
> experimentation. Maybe a 4 core CPU, with plenty of memory. Not sure
> what effect and bonuses you get with multiple cores, and what might be a
> minimum number of cores, and what might be optimum.
a minimum, but look for a system which can be expanded to more, and the
more the merrier. Also look at disk and interface expansion capability.
> As to disk storage. Are virtual devices set up for the guest OSs? Is
> multiple disks or partitions advisable? Or just have one large disk
> that gets virtual devices created for the guests?
As I said in my earlier reply to John M., multiple disks are the way to
go when dealing with virtual machines. At one point I had 3 disks in a
system - the system disk, the disk for VM clients, and another to hold
installation DVD images and backups. I also have a few external USB
disks for backups. Apart from installing the host OS, you will find you
don't really need a stack of DVDs any more - you download the boot
images and point your VM software at one of those when creating a client.
Also note that once you start creating VM clients, you will eat up disk
space like you would never have believed. Now that 2TB disks are common
and 3TB disks are around, you will probably find that you can get good
prices on 1.5TB or lesser capacity disks. Just what you already know
here.
And then there's the question of using an SSD for your system disk.
These are wickedly fast in comparison with traditional disks. :-)
--
Paul Sture
John E. Malmberg
Sep 24, 2012, 8:16:28 AM9/24/12
to
On 9/24/2012 5:51 AM, Paul Sture wrote:
>
> In the VirtualBox context, how do you ensure that the guest VMs are
> "configured with the paravirtualized I/O drivers"?
>
> I have VirtualBox Guest Additions installed on clients. Is that enough,
> or are there other drivers available?
There are some Microsoft Windows drivers from the KVM project hosted by
>
> In the VirtualBox context, how do you ensure that the guest VMs are
> "configured with the paravirtualized I/O drivers"?
>
> I have VirtualBox Guest Additions installed on clients. Is that enough,
> or are there other drivers available?
Red-Hat. There is a link in the Virtual Box documentation that is close
to where they are. I do not have time to look up the precise links
right now.
You also need to set the Emulated Network cards to be Paravirtualized IO
after downloading the drivers.
Tip: configure the VM with one emulated physical NIC, download the
paravirtualized driver, then switch the type of the VM and add the other
emulated NICs as paravirtualized. Then install the driver.
Most Linux distributions come with the paravirtualized drivers as part
of their kit if they are not in the host extensions.
>> I have not yet run the VUPOMETER tests on the emulated VMS systems, but
>> it might be interesting to see ratings from the various devices VMS has
>> been emulated on.
>
> Does anyone have a common version of VUPOMETER they agree on? Someone
> mentioned the Charon site earlier this year, but I couldn't find
> anything there.
http://vmshobbyist.com in the FreeAxp forum.
Paul Sture
Sep 24, 2012, 10:55:09 AM9/24/12
to
In article <nospam-4D9C1A....@news.chingola.ch>,
something to "speed up your Mac", but the Windows executable wants to
install AVG antivirus plus toolbar too. You don't need to install it,
but the installation program doesn't make that obvious.
I know people have to make a living, but I really don't like such
tactics.
--
Paul Sture
Paul Sture <nos...@sture.ch> wrote:
> In article <4LSdnRyVIdzTmcPN...@mchsi.com>,
> "John E. Malmberg" <wb8...@qsl.network> wrote:
>
> > I am also running syngery to remotely control the keyboard and mouse of
> > the Macbook, which is going to use some cycles.
>
> http://synergy-foss.org
>
> (site best viewed with Javascript turned off, to avoid invitations to
> download some other product completely. which will allegedly "Speed up
> your Mac")
>
> "Synergy lets you easily share your mouse and keyboard between multiple
> computers on your desk, and it's Free and Open Source. Just move your
> mouse off the edge of one computer's screen on to another. You can even
> share all of your clipboards. All you need is a network connection.
> Synergy is cross-platform (works on Windows, Mac OS X and Linux)."
>
> That's worth a look. Thanks for the tip.
>
Ouch. Not only is their website set up to persuade you download
> In article <4LSdnRyVIdzTmcPN...@mchsi.com>,
> "John E. Malmberg" <wb8...@qsl.network> wrote:
>
> > I am also running syngery to remotely control the keyboard and mouse of
> > the Macbook, which is going to use some cycles.
>
> http://synergy-foss.org
>
> (site best viewed with Javascript turned off, to avoid invitations to
> download some other product completely. which will allegedly "Speed up
> your Mac")
>
> "Synergy lets you easily share your mouse and keyboard between multiple
> computers on your desk, and it's Free and Open Source. Just move your
> mouse off the edge of one computer's screen on to another. You can even
> share all of your clipboards. All you need is a network connection.
> Synergy is cross-platform (works on Windows, Mac OS X and Linux)."
>
> That's worth a look. Thanks for the tip.
>
something to "speed up your Mac", but the Windows executable wants to
install AVG antivirus plus toolbar too. You don't need to install it,
but the installation program doesn't make that obvious.
I know people have to make a living, but I really don't like such
tactics.
--
Paul Sture
Anonymous
Sep 24, 2012, 12:25:15 PM9/24/12
to
Paul Sture <nos...@sture.ch> wrote:
> * I learned the hard way. My first VM capable PC came with Windows Home
> Premium, and I quickly learned that it contained too many restrictions
> for what I wanted to do. I very reluctantly forked out for an OEM copy
> of Windows 7 Ultimate, and felt I had been conned.
I have news for you buddy, whenever you pay for Windows you've been conned!
> * I learned the hard way. My first VM capable PC came with Windows Home
> Premium, and I quickly learned that it contained too many restrictions
> for what I wanted to do. I very reluctantly forked out for an OEM copy
> of Windows 7 Ultimate, and felt I had been conned.
David Froble
Sep 24, 2012, 4:52:19 PM9/24/12
to
Paul Sture wrote:
> Does anyone have a common version of VUPOMETER they agree on? Someone
> mentioned the Charon site earlier this year, but I couldn't find
> anything there.
I found the following laying around on one of my systems. I have no
> Does anyone have a common version of VUPOMETER they agree on? Someone
> mentioned the Charon site earlier this year, but I couldn't find
> anything there.
idea how (accurate) it might be, if accuracy can be used with anything
like this. A couple of lines may have been wrapped, should be easy to fix.
I've included some data from 2 of my systems. Couldn't be bothered to
fire up any of the EV6 systems to see how they do.
$! CALCULATE_VUPS:
$!
$ set noon
$ orig_privs = f$setprv("ALTPRI")
$ process_priority = f$getjpi(0,"PRIB")
$ cpu_multiplier = 10 ! VAX = 10 - Alpha/AXP = 40
$ cpu_round_add = 1 ! VAX = 1 - Alpha/AXP = 9
$ cpu_round_divide = cpu_round_add + 1
$ init_counter = cpu_multiplier * 525
$ init_loop_maximum = 205
$ start_cputime = f$getjpi(0,"CPUTIM")
$ loop_index = 0
$ 10$:
$ loop_index = loop_index + 1
$ if loop_index .ne. init_loop_maximum then goto 10$
$ end_cputime = f$getjpi(0,"CPUTIM")
$ init_vups = ((init_counter / (end_cputime - start_cputime) + -
cpu_round_add) / cpu_round_divide) * cpu_round_divide
$ loop_maximum = (init_vups * init_loop_maximum) / 10
$ base_counter = (init_counter * init_vups) / 10
$ vups = 0
$ times_through_loop = 0
$ 20$:
$ start_cputime = f$getjpi(0,"CPUTIM")
$ loop_index = 0
$ 30$:
$ loop_index = loop_index + 1
$ if loop_index .ne. loop_maximum then goto 30$
$ end_cputime = f$getjpi(0,"CPUTIM")
$ new_vups = ((base_counter / (end_cputime - start_cputime) + -
cpu_round_add) / cpu_round_divide) * cpu_round_divide
$ if new_vups .eq. vups then goto 40$
$ vups = new_vups
$ times_through_loop = times_through_loop + 1
$ if times_through_loop .le. 5 then goto 20$
$ 40$:
$ new_privs = f$setprv(orig_privs)
$ set message /nofacility/noidentification/noseverity/notext
$! ASSIGN/SYSTEM/EXEC 'vups' MACHINE_VUPS_RATING
$ set message /facility/identification/severity/text
$ write sys$output "Approximate System VUPs Rating : ", -
vups / 10,".", vups - ((vups / 10) * 10)
$ exit
---------------------------------------------------------
$ show lic/cha
VMS/LMF Charge Information for node DFE90A
This is a VAXstation 4000-90A, hardware model type 475
$ @calcvups
Approximate System VUPs Rating : 27.0
---------------------------------------------------------
AS800> show lic/cha
VMS/LMF Charge Information for node AS800
This is a AlphaServer 800 5/500, hardware model type 1585
AS800> @calcvups
Approximate System VUPs Rating : 141.8
---------------------------------------------------------
Might be interesting to see the numbers from some other systems ...
Jan-Erik Soderholm
Sep 24, 2012, 5:40:03 PM9/24/12
to
$ sh lic/cha
VMS/LMF Charge Information for node OSSBY1
This is a AlphaServer DS25, hardware model type 1994
Using your routine above :
$ @ calc_vups.com
Approximate System VUPs Rating : 491.4
Using a slighly modified routine :
$ @CALCULATE_VUP.COM
INFO: Preventing endless loop (10$) on fast CPUs
Approximate System VUPs Rating : 509.9 ( min: 508.8 max: 512.6 )
$
The updated version might be one found here :
http://www.openvmshobbyist.com/forum/viewthread.php?forum_id=163&thread_id=842
Single Stage to Orbit
Sep 24, 2012, 5:11:00 PM9/24/12
to
On Mon, 2012-09-24 at 16:52 -0400, David Froble wrote:
>
> Might be interesting to see the numbers from some other systems ...
On my SimH VAX simulation running VMS 7.3, on a Pentium IV 2.8GHz
>
> Might be interesting to see the numbers from some other systems ...
laptop, I get the following;
> @vups.com
Approximate System VUPs Rating : 7.0
--
Tactical Nuclear Kittens
Mike Rechtman
Sep 25, 2012, 12:25:10 AM9/25/12
to
VMS/LMF Charge Information for node MIKE
This is a AlphaStation 200 4/100, hardware model type 1156
->COMFILES_$@calcvups
Approximate System VUPs Rating : 70.0
--
Mike R.
Home: http://alpha.mike-r.com/
QOTD: http://alpha.mike-r.com/php/qotd.php
No Micro$oft products were used in the URLs above, or in preparing this message.
Recommended reading: http://www.catb.org/~esr/faqs/smart-questions.html#before
--- news://freenews.netfront.net/ - complaints: ne...@netfront.net ---
Artem Alimarin
Oct 17, 2012, 12:19:48 PM10/17/12
to
AlphaVM is not ported to OSX yet. This is planned but has much lower priority than other things that have commercial justification ;) Anyway, this is the next candidate to port after I get myself a MAC.
On Monday, September 17, 2012 5:44:57 AM UTC+2, Sum1 wrote:
> Downloaded the Linux version for a play - would like to run VMS on OSX
>
> that is more current than 7.3-2 (SimH emulating Alpha isn't available
>
> yet…)
>
>
>
> Has anyone got AlphaVM working on OSX and if so, could you please share
>
> the details?
>
>
>
> Cheers
>
>
>
> PS Many thanks for those of you who helped with my previous dd problem.
>
> Indeed I had imaged too much - I imaged the whole CD rather than just
>
> the VMS partition, once corrected the iso worked fine :)
On Monday, September 17, 2012 5:44:57 AM UTC+2, Sum1 wrote:
> Downloaded the Linux version for a play - would like to run VMS on OSX
>
> that is more current than 7.3-2 (SimH emulating Alpha isn't available
>
> yet…)
>
>
>
> Has anyone got AlphaVM working on OSX and if so, could you please share
>
> the details?
>
>
>
> Cheers
>
>
>
> PS Many thanks for those of you who helped with my previous dd problem.
>
> Indeed I had imaged too much - I imaged the whole CD rather than just
>
> the VMS partition, once corrected the iso worked fine :)
Artem Alimarin
Oct 17, 2012, 12:24:00 PM10/17/12
to
I am afraid you are confusing AlphaVM with vtAlpha. AlphaVM was never advertised as running on "bare metal".
On Monday, September 17, 2012 2:31:36 PM UTC+2, Stephen Hoffman wrote:
[snip]
> IIRC, the EmuVM folks had some seriously confusing advertising with
>
> their "bare-metal" stuff a while back.
>
[/snip]
On Monday, September 17, 2012 2:31:36 PM UTC+2, Stephen Hoffman wrote:
[snip]
> IIRC, the EmuVM folks had some seriously confusing advertising with
>
> their "bare-metal" stuff a while back.
>
[/snip]
Stephen Hoffman
Oct 17, 2012, 12:39:03 PM10/17/12
to
markw...@googlemail.com
Feb 26, 2013, 6:43:57 AM2/26/13
to
I have got the Alpha ES40.org simulator (the 0.18+) latest version compiled and running openVMS 8.4 directly on a Macbook Air. It hogs the CPU like mad but it runs. You have to run it as root which I don't really like either. Not tried the networking - just got it to a stage it accepts License PAKs and can generate a new disk image and boot that. It is a pain to get the SRM files and an image into place, but it works.
www.es40.org has details and how to do it in OSX.
http://www.es40.org/OS_X is the instruction listing- you need to download Xcode and the additional developer tools for Mac OSX and create your own version which ends up in usr/local/bin
Also get the latest latest snapshot from the CVS on es40.org
I have tried FreeAXP under Parallels Desktop 8 under windows 7 which is an easy install and easier to control frankly.
www.es40.org has details and how to do it in OSX.
http://www.es40.org/OS_X is the instruction listing- you need to download Xcode and the additional developer tools for Mac OSX and create your own version which ends up in usr/local/bin
Also get the latest latest snapshot from the CVS on es40.org
I have tried FreeAXP under Parallels Desktop 8 under windows 7 which is an easy install and easier to control frankly.
Artem Alimarin
May 22, 2013, 11:09:12 AM5/22/13
to
We have published AlphaVM benchmarks at emuvm.com/benchmarks.php. We are getting close to fastest real Alpha's. VUPs rating depends a lot on the OpenVMS version used to run the benchmark. Therefore we measured several versions.
AlphaVM-Pro on E3-1270v2, 3.5GHz gives the following values:
OpenVMS 7.1-2 - 560.0
OpenVMS 7.3-1 - 519.0
OpenVMS 7.3-2 - 439.9
OpenVMS 8.3 - 503.2
OpenVMS 8.4 - 491.9
AlphaVM-Pro on E3-1270v2, 3.5GHz gives the following values:
OpenVMS 7.1-2 - 560.0
OpenVMS 7.3-1 - 519.0
OpenVMS 7.3-2 - 439.9
OpenVMS 8.3 - 503.2
OpenVMS 8.4 - 491.9
0 new messages