- Replace the need of provide Engine API URL
- Remove the need of copy/past certificate
- Auto-dectect if Engine CA cert is installed
- Check if engine domain provided is responsive
- User cannot fix wrong oVirt user once provided
- Replace static strings in the code from oVirt to Engine
- Do not allow users to select uninitialized DC
- Credentials - set max login attempts
- ovirt: DNS VIP removal
- Update explain for installer
Signed-off-by: Douglas Schilling Landgraf <dougsland@redhat.com>
Those items are now part of MachinePool
- instance type
- cpu (cores, sockets)
- memory
- os disk size
- vm type
With those we have controll on the way create worker/nodes,
and we have proper defaults set in the machine pool definitions, one for
the pkg/asset/machines/ master.go and worker.go.
Default master pool:
cpus: 4
mem: 16 GiB
os disk: 120 GB
VM type: high_performance
Default worker pool:
cpus: 4
mem: 16 GiB
os disk: 120 GB
VM type: server
A compute pool snippet from the install-config.yaml may look like that:
```yaml
compute:
- architecture: amd64
hyperthreading: Enabled
name: worker
platform:
ovirt:
cpu:
cores: 8
sockets: 1
instanceTypeID:
memoryMB: 65536
osDisk:
sizeGB: 120
vmType: high_performance
replicas: 3
```
Terraform now uses those values, and all the defaults in tf files are
now deleted in favour of the machine-config ones.
To support `osDisk` size which is differnt than the template
master/main.tf has this block:
block_device {
interface = "virtio_scsi"
size = var.ovirt_master_os_disk_size_gb
}
If needed TF will extend the disk size of the VM.
Cluster-api-provider-ovirt also extends the disk of a VM if the
definition in the machine spec is differnt than the template.
Signed-off-by: Roy Golan <rgolan@redhat.com>
