Discussion:
OpenVMS x86 and qemu/kvm - boot problems.
(too old to reply)
Timothy Stark
2023-04-12 12:49:53 UTC
Permalink
I watched a video clip on YouTube about installing OpenVMS on x86_64. I
followed instructions from that. It did not work well. He did not show how
to set bridge br0 for kvm.

It crashed during booting. I got two error messages said unable determine
serial number in HWPB block and bug check message said that high IPL during
pagefault exception. I think that I got wrong UEFI firmware from
/usr/share/qemu. Does anyone have good UEFI firmware?

I tried to configure br0 but I got error said that they can't find bridge
file (-net bridge,br0). Does anyone have configuration info for setting
br0 device on Ubuntu 22.10?

I successfully installed OpenVMS on virtual box under windows 11. It
recognized 82450EM and assigned EIA0 to that. However it can't access the
Internet so that I can't upload layered products into. Does anyone have
any good network configuration for VirtualBox 7.x?
Dennis Boone
2023-04-12 15:51:47 UTC
Permalink
Post by Timothy Stark
It crashed during booting. I got two error messages said unable determine
serial number in HWPB block and bug check message said that high IPL during
pagefault exception. I think that I got wrong UEFI firmware from
/usr/share/qemu. Does anyone have good UEFI firmware?
I have 2022.11-6 of the ovmf package and qemu 7.2.on a Debian Sid box.

ozymandias 1643 # sha256sum /usr/share/ovmf/OVMF.fd
d996e65015d52ad21ec4014c4f6f80038dbfd63463106e39eb66b81282ab8dbf /usr/share/ovmf/OVMF.fd

-rw-r--r-- 1 root root 2097152 mar 5 15:36 OVMF.fd

Make sure you're not trying to boot in secure mode.

I get the serial number message too, but it doesn't seem to prevent
booting and installation.

I can reliably induce a crash (high IPL, page fault) by trying to
mount the ISO image. I'm also getting random crashes I haven't
diagnosed yet. I.e. ymmv.

De
issinoho
2023-04-20 16:11:46 UTC
Permalink
Post by Dennis Boone
Post by Timothy Stark
It crashed during booting. I got two error messages said unable determine
serial number in HWPB block and bug check message said that high IPL during
pagefault exception. I think that I got wrong UEFI firmware from
/usr/share/qemu. Does anyone have good UEFI firmware?
I have 2022.11-6 of the ovmf package and qemu 7.2.on a Debian Sid box.
ozymandias 1643 # sha256sum /usr/share/ovmf/OVMF.fd
d996e65015d52ad21ec4014c4f6f80038dbfd63463106e39eb66b81282ab8dbf /usr/share/ovmf/OVMF.fd
-rw-r--r-- 1 root root 2097152 mar 5 15:36 OVMF.fd
Make sure you're not trying to boot in secure mode.
I get the serial number message too, but it doesn't seem to prevent
booting and installation.
I can reliably induce a crash (high IPL, page fault) by trying to
mount the ISO image. I'm also getting random crashes I haven't
diagnosed yet. I.e. ymmv.
De
Today I've finally got this working on QEMU/KVM using the virt-manager GUI rather than an enormous command line of switches. Everything work perfectly, albeit I'm running in NAT mode. Bridging looks to be a world of pain so I'm next going to look at NAT forwarding through the Ubuntu host - will report back.

X86VMS::sh cpu

System: X86VMS, QEMU Standard PC (Q35 + ICH9, 2009)
Chris Townley
2023-04-20 18:33:41 UTC
Permalink
Post by issinoho
Post by Dennis Boone
Post by Timothy Stark
It crashed during booting. I got two error messages said unable determine
serial number in HWPB block and bug check message said that high IPL during
pagefault exception. I think that I got wrong UEFI firmware from
/usr/share/qemu. Does anyone have good UEFI firmware?
I have 2022.11-6 of the ovmf package and qemu 7.2.on a Debian Sid box.
ozymandias 1643 # sha256sum /usr/share/ovmf/OVMF.fd
d996e65015d52ad21ec4014c4f6f80038dbfd63463106e39eb66b81282ab8dbf /usr/share/ovmf/OVMF.fd
-rw-r--r-- 1 root root 2097152 mar 5 15:36 OVMF.fd
Make sure you're not trying to boot in secure mode.
I get the serial number message too, but it doesn't seem to prevent
booting and installation.
I can reliably induce a crash (high IPL, page fault) by trying to
mount the ISO image. I'm also getting random crashes I haven't
diagnosed yet. I.e. ymmv.
De
Today I've finally got this working on QEMU/KVM using the virt-manager GUI rather than an enormous command line of switches. Everything work perfectly, albeit I'm running in NAT mode. Bridging looks to be a world of pain so I'm next going to look at NAT forwarding through the Ubuntu host - will report back.
X86VMS::sh cpu
System: X86VMS, QEMU Standard PC (Q35 + ICH9, 2009)
Much easier in the GUI, bit it needs too many tweaks.

A bridge network is pretty easy to set up. I have a second NIC (USB)
that I have bridged to. Setup took me a while, but when I recreated my
KVM/QEMU VM I got it working fine just by selecting Bridged, and
choosing the br0 device I had previously created.

Sadly I didn't note the exact way I created a br0 device on the host,
but plenty of examples on the net.

Have you managed to get he serial console accessible over ssh? So far I
have been stuck with the Virtual Machine builder serial console, which
is pretty carp.
--
Chris
issinoho
2023-04-20 21:27:05 UTC
Permalink
Post by Chris Townley
Post by issinoho
Post by Dennis Boone
Post by Timothy Stark
It crashed during booting. I got two error messages said unable determine
serial number in HWPB block and bug check message said that high IPL during
pagefault exception. I think that I got wrong UEFI firmware from
/usr/share/qemu. Does anyone have good UEFI firmware?
I have 2022.11-6 of the ovmf package and qemu 7.2.on a Debian Sid box.
ozymandias 1643 # sha256sum /usr/share/ovmf/OVMF.fd
d996e65015d52ad21ec4014c4f6f80038dbfd63463106e39eb66b81282ab8dbf /usr/share/ovmf/OVMF.fd
-rw-r--r-- 1 root root 2097152 mar 5 15:36 OVMF.fd
Make sure you're not trying to boot in secure mode.
I get the serial number message too, but it doesn't seem to prevent
booting and installation.
I can reliably induce a crash (high IPL, page fault) by trying to
mount the ISO image. I'm also getting random crashes I haven't
diagnosed yet. I.e. ymmv.
De
Today I've finally got this working on QEMU/KVM using the virt-manager GUI rather than an enormous command line of switches. Everything work perfectly, albeit I'm running in NAT mode. Bridging looks to be a world of pain so I'm next going to look at NAT forwarding through the Ubuntu host - will report back.
X86VMS::sh cpu
System: X86VMS, QEMU Standard PC (Q35 + ICH9, 2009)
Much easier in the GUI, bit it needs too many tweaks.
A bridge network is pretty easy to set up. I have a second NIC (USB)
that I have bridged to. Setup took me a while, but when I recreated my
KVM/QEMU VM I got it working fine just by selecting Bridged, and
choosing the br0 device I had previously created.
Sadly I didn't note the exact way I created a br0 device on the host,
but plenty of examples on the net.
Have you managed to get he serial console accessible over ssh? So far I
have been stuck with the Virtual Machine builder serial console, which
is pretty carp.
--
Chris
On NAT port forwarding, this works beautifully, https://blog.wirelessmoves.com/2022/04/bare-metal-cloud-part-2-kvm-and-nat-port-forwarding.html

On console, I'm using 'virsh console vm-name' just enough to get the OpenVMS networking up then using SSH from that point on, initially from the host, but once you use that ^^^ script from anywhere on your network.
issinoho
2023-04-21 09:48:17 UTC
Permalink
Post by issinoho
Post by Chris Townley
Post by issinoho
Post by Dennis Boone
Post by Timothy Stark
It crashed during booting. I got two error messages said unable determine
serial number in HWPB block and bug check message said that high IPL during
pagefault exception. I think that I got wrong UEFI firmware from
/usr/share/qemu. Does anyone have good UEFI firmware?
I have 2022.11-6 of the ovmf package and qemu 7.2.on a Debian Sid box.
ozymandias 1643 # sha256sum /usr/share/ovmf/OVMF.fd
d996e65015d52ad21ec4014c4f6f80038dbfd63463106e39eb66b81282ab8dbf /usr/share/ovmf/OVMF.fd
-rw-r--r-- 1 root root 2097152 mar 5 15:36 OVMF.fd
Make sure you're not trying to boot in secure mode.
I get the serial number message too, but it doesn't seem to prevent
booting and installation.
I can reliably induce a crash (high IPL, page fault) by trying to
mount the ISO image. I'm also getting random crashes I haven't
diagnosed yet. I.e. ymmv.
De
Today I've finally got this working on QEMU/KVM using the virt-manager GUI rather than an enormous command line of switches. Everything work perfectly, albeit I'm running in NAT mode. Bridging looks to be a world of pain so I'm next going to look at NAT forwarding through the Ubuntu host - will report back.
X86VMS::sh cpu
System: X86VMS, QEMU Standard PC (Q35 + ICH9, 2009)
Much easier in the GUI, bit it needs too many tweaks.
A bridge network is pretty easy to set up. I have a second NIC (USB)
that I have bridged to. Setup took me a while, but when I recreated my
KVM/QEMU VM I got it working fine just by selecting Bridged, and
choosing the br0 device I had previously created.
Sadly I didn't note the exact way I created a br0 device on the host,
but plenty of examples on the net.
Have you managed to get he serial console accessible over ssh? So far I
have been stuck with the Virtual Machine builder serial console, which
is pretty carp.
--
Chris
On NAT port forwarding, this works beautifully, https://blog.wirelessmoves.com/2022/04/bare-metal-cloud-part-2-kvm-and-nat-port-forwarding.html
On console, I'm using 'virsh console vm-name' just enough to get the OpenVMS networking up then using SSH from that point on, initially from the host, but once you use that ^^^ script from anywhere on your network.
This is starting to take shape nicely. VM now autostarting on Ubuntu with persistent firewall rules port forwarding SSH and HTTP, and X sessions tunneling nicely back to a Windows desktop.
John H. Reinhardt
2024-05-28 20:40:28 UTC
Permalink
Have you managed to get he serial console accessible over ssh? So far I have been stuck with the Virtual Machine builder serial console, which is pretty carp.
I've got a ProxMox V8.2.2 instance running now. It boots the OpenVMS V9.2-2 ISO but for the life of me I can't get ssh/telnet setup to the serial console. The documentation for this is quite obtuse. As with you I have the serial console connected in the ProxMox console but I want to be able to telnet or ssh to it from other places than where the web gui is running.
--
John H. Reinhardt
Chris Townley
2024-05-28 22:58:05 UTC
Permalink
Post by Chris Townley
Have you managed to get he serial console accessible over ssh? So far
I have been stuck with the Virtual Machine builder serial console,
which is pretty carp.
I've got a ProxMox V8.2.2 instance running now.  It boots the OpenVMS
V9.2-2 ISO but for the life of me I can't get ssh/telnet setup to the
serial console.  The documentation for this is quite obtuse.  As with
you I have the serial console connected in the ProxMox console but I
want to be able to telnet or ssh to it from other places than where the
web gui is running.
I never managed to get any serial terminal running, but I am on KVM/QEMU
under Ubuntu - I don't know anything about Proxmox.

I got a PuTTy instance running on the host, then issued:
$ virsh console <VM_name>

That works well as a console.

When DECNET, TCP and SSH installed - my network is bridged to the host's
second NIC, I has no problem using ssh from my Windoze workstation
--
Chris
John H. Reinhardt
2024-05-29 02:32:08 UTC
Permalink
Have you managed to get he serial console accessible over ssh? So far I have been stuck with the Virtual Machine builder serial console, which is pretty carp.
I've got a ProxMox V8.2.2 instance running now.  It boots the OpenVMS V9.2-2 ISO but for the life of me I can't get ssh/telnet setup to the serial console.  The documentation for this is quite obtuse.  As with you I have the serial console connected in the ProxMox console but I want to be able to telnet or ssh to it from other places than where the web gui is running.
I never managed to get any serial terminal running, but I am on KVM/QEMU under Ubuntu - I don't know anything about Proxmox.
ProxMox is based on KVM, but it's got cluster aware parts. There is a distributed, shared file system that it interfaces to called Ceph (as do others). The aim is to have shared virtual disks for VMSClusters with common shared drives. Not sure if that's possible yet, but I think all the parts are there.

ProxMox - https://www.proxmox.com/en/
Ceph - https://ceph.io/en/
$ virsh console <VM_name>
That works well as a console.
When DECNET, TCP and SSH installed - my network is bridged to the host's second NIC, I has no problem using ssh from my Windoze workstation
--
John H. Reinhardt
Michael Brown
2024-06-07 02:42:38 UTC
Permalink
I've got a ProxMox V8.2.2 instance running now.  It boots the OpenVMS
V9.2-2 ISO .......
Is there a secret to getting the network device working under proxmox ?
I used e1000 as the device model but during bootup it sends the cpu
usage up to 100% and just freezes. Using other device models some will
seemingly fail immediately and the bootup goes ahead but you end up with
openvms running but with no network device available.

Have you seen anything like this happening before.
John H. Reinhardt
2024-06-09 04:34:30 UTC
Permalink
Post by Michael Brown
I've got a ProxMox V8.2.2 instance running now.  It boots the OpenVMS V9.2-2 ISO .......
Is there a secret to getting the network device working under proxmox ?
I used e1000 as the device model but during bootup it sends the cpu usage up to 100% and just freezes. Using other device models some will seemingly fail immediately and the bootup goes ahead but you end up with openvms running but with no network device available.
Have you seen anything like this happening before.
Go back to using the e1000 device. After you create the VM, but before you boot it, edit the VM config file /etc/pve/qemu-server/<vid>.conf

Add "args: -machine hpet=off" if it's not already there. If it wasn't there, then that should fix the problem.

This is my current test config file:

pve011:/etc/pve/qemu-server# cat 100.conf
args: -machine hpet=off
balloon: 0
bios: ovmf
boot: order=ide2;sata0;sata1;net0
cores: 2
cpu: host
efidisk0: local-lvm:vm-100-disk-0,efitype=4m,size=4M
ide2: local:iso/X860922OE.ISO,media=cdrom,size=1517408K
machine: q35
memory: 8192
meta: creation-qemu=8.1.5,ctime=1717891692
name: openvms-test
net0: e1000=BC:24:11:A5:FB:1A,bridge=vmbr0,firewall=1,tag=16
net1: e1000=BC:24:11:0D:4F:69,bridge=vmbr0,firewall=1
net2: e1000=BC:24:11:0C:52:86,bridge=vmbr0,firewall=1,tag=20
numa: 0
ostype: other
sata0: local-lvm:vm-100-disk-2,discard=on,size=16G,ssd=1
sata1: local-lvm:vm-100-disk-1,discard=on,size=16G,ssd=1
scsihw: virtio-scsi-pci
serial0: socket
smbios1: uuid=1fa8bf23-cff3-4746-91cb-d9fed4ea99c5
sockets: 1
vmgenid: ce4c0615-9dc0-46e5-8f5c-5a4eca5b7647
--
John H. Reinhardt
Michael Brown
2024-06-12 04:20:30 UTC
Permalink
Post by Michael Brown
I've got a ProxMox V8.2.2 instance running now.  It boots the OpenVMS
V9.2-2 ISO .......
Is there a secret to getting the network device working under proxmox ?
Go back to using the e1000 device.  After you create the VM, but before
you boot it, edit the VM config file /etc/pve/qemu-server/<vid>.conf
Add "args: -machine hpet=off" if it's not already there.  If it wasn't
there, then that should fix the problem.
Yes! that was the secret. vms boots up now and is on the net. Thank you
muchly.

Loading...