Question about where we'll be able to run OpenVMS x86
-
Topic author - Visitor
- Posts: 2
- Joined: Tue May 09, 2023 5:18 pm
- Reputation: 0
- Status: Offline
Question about where we'll be able to run OpenVMS x86
So I'm completely new to OpenVMS other than a bit of fiddling with the Alpha version in a VM. I get that the current x86 field test is mostly aimed at consumer virtualization software, but I'm curious about where I can expect to be able to run anything I develop on the field test in the future.
Primarily, I was wondering - can I expect that I will be able to run it on bare metal on the average desktop computer or as a custom ISO on a service like Digital Ocean or Vultr?
Primarily, I was wondering - can I expect that I will be able to run it on bare metal on the average desktop computer or as a custom ISO on a service like Digital Ocean or Vultr?
Re: Question about where we'll be able to run OpenVMS x86
VMS is VMS is VMS.
So what you get now will be the same when the missing bits has been complete.
And it will run *supported* on whatever is specified in the road-mapp and other documents.
So what you get now will be the same when the missing bits has been complete.
And it will run *supported* on whatever is specified in the road-mapp and other documents.
Re: Question about where we'll be able to run OpenVMS x86
There is already a discussion thread about running on dedicated hardware, there are some reports.
viewtopic.php?f=21&t=8621
As always - the challenges are related to the hardware being used, as I can see most related to the storage controller.
That's why the current most promising approach is to use a hypervisor, as those are offering LSI like controllers.
viewtopic.php?f=21&t=8621
As always - the challenges are related to the hardware being used, as I can see most related to the storage controller.
That's why the current most promising approach is to use a hypervisor, as those are offering LSI like controllers.
Re: Question about where we'll be able to run OpenVMS x86
Many of the cloud providers such as Amazon use the KVM hypervisor under the hood for their cloud instances, so it should be possible to deploy VMS x86 there. Perhaps the most notable exception is Microsoft who use Hyper-V, which is currently not supported by VMS.
-
Topic author - Visitor
- Posts: 2
- Joined: Tue May 09, 2023 5:18 pm
- Reputation: 0
- Status: Offline
Re: Question about where we'll be able to run OpenVMS x86
Thanks for the responses. I'll have to see if deploying it on a cloud service will work once I'm more familiar with it.
-
- Member
- Posts: 5
- Joined: Thu Nov 03, 2022 10:45 am
- Reputation: 0
- Status: Offline
Re: Question about where we'll be able to run OpenVMS x86
I think it will be really interesting to see what happens once an Amazon EC2 AMI for VMS is in place, where anyone can come along and spin up VMS, and where VSI can monetize VMS based on how long instances are run by AWS customers. Admittedly, I don't have any idea of how much trouble it will be to make VMS work that way.
It seem that having easily launched VMS cloud instances will provide access to a lot more people. Hopefully the cloud providers will keep the burden low enough so that VSI can make a decent profit from customers running in the cloud.
It seem that having easily launched VMS cloud instances will provide access to a lot more people. Hopefully the cloud providers will keep the burden low enough so that VSI can make a decent profit from customers running in the cloud.
-
- VSI Expert
- Contributor
- Posts: 18
- Joined: Mon Aug 09, 2021 9:01 am
- Reputation: 0
- Status: Offline
Re: Question about where we'll be able to run OpenVMS x86
Back to the original question....
We have no current plans to support x86 VMS on bare metal. Would we if someone offered us a very large pile of money? Maybe, but it would have to be a very unique situation.
Clair
We have no current plans to support x86 VMS on bare metal. Would we if someone offered us a very large pile of money? Maybe, but it would have to be a very unique situation.
Clair
-
- Master
- Posts: 475
- Joined: Fri Apr 17, 2020 7:31 pm
- Reputation: 0
- Location: Rhode Island, USA
- Status: Offline
- Contact:
Re: Question about where we'll be able to run OpenVMS x86
I can certainly understand that running on VM is the priority right now. Most customer will want to run VMS on VM's (most likely ESXi for on-prem and for cloud whatever their favorite public cloud vendor happens to use). And physical x86-64 servers are much worse in number of possible combinations than virtual servers.
But at least some VMS users are very conservative. The "anything less than 25 years old is new stuff" crowd. They may not be comfortable with virtualization. And telling them that they are wrong will not work - as the old saying goes "The customer is always right even when wrong".
When things start to settle and VSI get some available resources, then maybe it would make sense to certify a single physical server. One server with one combo of CPU, NIC's, controllers and storage. Anyone that want physical will need to buy one of those.
One such for VMS 9.x, another newer one for 10.x, another for 11.x and so on. I don't know the expected interval for new major VMS versions, but if we say every 5 years, then it would mean change of supported physical server every 5 years. That server would be pretty old at the end of the period, but if someone don't want that "new" technology called virtualization, then they should be able to live with a 5 year old server.
-
- Active Contributor
- Posts: 31
- Joined: Sat Apr 22, 2023 3:56 am
- Reputation: 0
- Status: Offline
Re: Question about where we'll be able to run OpenVMS x86
The nice thing about a lot of this hardware support though, is that as say, even just VM platform support is improved, it has a "bleed through" effect of supporting more and varied scenarios.
A lot of the hardware related issues i'm seeing crash wise/boot will be resolved over time most likely just through the work of people testing and documenting system parameters, improvement in ACPI support/testing in various configurations, and as noted above, additional VM platforms that present different configurations.
As long as that fix is proper and isn't a "if vmware 8, assume these values" type fix anyway - that'd be a pain point for any newer virt environments or even going-forward improvements in hypervisors as the years tick on.
That is, of course, not to say VSI will *support* your configuration, but over time more and more variations will *probably* start working - IE disregard this unknown or better memory mapping/detection as edge cases are discovered, etc. So you'll still have your laundry list of unsupported devices, but more boot scenarios will be possible.
A lot of the hardware related issues i'm seeing crash wise/boot will be resolved over time most likely just through the work of people testing and documenting system parameters, improvement in ACPI support/testing in various configurations, and as noted above, additional VM platforms that present different configurations.
As long as that fix is proper and isn't a "if vmware 8, assume these values" type fix anyway - that'd be a pain point for any newer virt environments or even going-forward improvements in hypervisors as the years tick on.
That is, of course, not to say VSI will *support* your configuration, but over time more and more variations will *probably* start working - IE disregard this unknown or better memory mapping/detection as edge cases are discovered, etc. So you'll still have your laundry list of unsupported devices, but more boot scenarios will be possible.
Last edited by gdwnldsksc on Wed May 24, 2023 7:04 pm, edited 2 times in total.
Re: Question about where we'll be able to run OpenVMS x86
On VMS platforms, I'm thinking another one that would be very interesting is 64 bit ARM. That's arguably the way of the future, and certainly a great way to run software on very power-efficient devices.