Unable to launch VM on Red Hat Enterprise Linux release 9.0 (Plow)

[tosin2013]

I ran the following command on RHEL 9 and I am receiving the error below.

sudo kcli create vm -p ztpfwjumpbox jumpbox --wait
➜ sudo kcli create vm -p ztpfwjumpbox jumpbox --wait
[sudo] password for admin: 
Deploying vm jumpbox from profile ztpfwjumpbox...
xorriso 1.5.4 : RockRidge filesystem manipulator, libburnia project.

Got storage volume 'jumpbox.ISO' exists already when creating iso
Traceback (most recent call last):
  File "/bin/kcli", line 33, in <module>
    sys.exit(load_entry_point('kcli==99.0', 'console_scripts', 'kcli')())
  File "/usr/lib/python3.9/site-packages/kvirt/cli.py", line 5058, in cli
    args.func(args)
  File "/usr/lib/python3.9/site-packages/kvirt/cli.py", line 1334, in create_vm
    result = config.create_vm(name, profile, overrides=overrides, customprofile=customprofile,
  File "/usr/lib/python3.9/site-packages/kvirt/config.py", line 1009, in create_vm
    result = k.create(name=name, virttype=virttype, plan=plan, profile=profilename, flavor=flavor,
  File "/usr/lib/python3.9/site-packages/kvirt/providers/kvm/__init__.py", line 1266, in create
    conn.defineXML(vmxml)
  File "/usr/lib64/python3.9/site-packages/libvirt.py", line 4414, in defineXML
    raise libvirtError('virDomainDefineXML() failed')
libvirt.libvirtError: unsupported configuration: spice graphics are not supported with this QEMU

Profile Example

ztpfwjumpbox:
 image: 'CentOS-Stream-GenericCloud-9-20220718.0.x86_64.qcow2'
 numcpus: 4
 memory: 8192
 disks:
  - size: 50
 reservedns: true
 nets:
  - name: qubinet
  - name: bare-net
 cmds: 
  - echoCHANGEME| passwd --stdin root
  - useradd admin
  - usermod -aG wheel admin
  - echo CHANGEME | passwd --stdin admin
  - dnf groupinstall "Server with GUI" -y
  - dnf install epel-release -y
  - dnf install xrdp  -y
  - systemctl enable xrdp --now
  - firewall-cmd --permanent --add-port=3389/tcp 
  - firewall-cmd --reload

[tosin2013]

$ cat /usr/lib/python3.9/site-packages/kvirt/providers/kvm/__init__.py
 296         if vnc:
 297             display = 'vnc'
 298         else:
 299             display = 'spice'
 300         volumes = {}

[karmab]

do you mind validating that a vm created with vnc does launch fine? just use vnc: true in your profile for that matter

[tosin2013]

When I use a bridge network it gets stuck in this loop. but VNC does work in the cockpit UI.

sudo kcli create vm -p centos9stream testvm
Deploying vm testvm from profile centos9stream...
xorriso 1.5.4 : RockRidge filesystem manipulator, libburnia project.

Creating dns entry for testvm in network qubinet
Waiting 5 seconds to grab ip...
Waiting 5 seconds to grab ip...
Waiting 5 seconds to grab ip...
Waiting 5 seconds to grab ip...
Waiting 5 seconds to grab ip...
Waiting 5 seconds to grab ip...
Waiting 5 seconds to grab ip...
Waiting 5 seconds to grab ip...
Waiting 5 seconds to grab ip...
Waiting 5 seconds to grab ip...
Waiting 5 seconds to grab ip...
Waiting 5 seconds to grab ip...
Waiting 5 seconds to grab ip...

The same issue occurs when i use nat bare-net

sudo kcli create vm -p centos9stream testvm
Deploying vm testvm from profile centos9stream...
xorriso 1.5.4 : RockRidge filesystem manipulator, libburnia project.

Creating dns entry for testvm in network qubinet
Waiting 5 seconds to grab ip...
Waiting 5 seconds to grab ip...
Waiting 5 seconds to grab ip...
Waiting 5 seconds to grab ip...
Waiting 5 seconds to grab ip...
Waiting 5 seconds to grab ip...
Waiting 5 seconds to grab ip...
Waiting 5 seconds to grab ip...
Waiting 5 seconds to grab ip...
Waiting 5 seconds to grab ip...
Waiting 5 seconds to grab ip...
Waiting 5 seconds to grab ip...
Waiting 5 seconds to grab ip...
Waiting 5 seconds to grab ip...
Waiting 5 seconds to grab IP...

[karmab]

sorry, been afk last entire week, i will be back online on wednesday and will test rhel9 hv

[karmab]

so indeed spice is getting removed from rhel9, as indicated in https://access.redhat.com/articles/442543 using vnc does work as expected though so i need to decide whether:

• we default to vnc from now on
• we force vnc if we can somehow figure out the hypervisor is rhel9

[karmab]

fixed in https://github.com/karmab/kcli/commit/d399261800adb59fdfed37a97ee08b8690ed6c66 where we default to vnc

(a following commit made sure to handle vnc clients from macosx by leveraging built in vnc client, at the cost of using a hidden vnc password in this special case since it's required by said client)