Recent version of CDI will garbage collect completed DataVolume. For VM uses DataVolume, virtink waits its DataVolume import to be completed before starting the VM. But with default CDI deployment, DataVolume will be GCed once completed. Hence virtink fails deploy VM with build VM Pod: DataVolume.cdi.kubevirt.io "cdi-volume" not found error.
Perhaps it's worth mentioning this in CDI VM example. Tuning .spec.config.dataVolumeTTLSeconds of cdi.kubevirt.io/v1beta1 works.(or adapting virtink to CDI's GC action?)
Recent version of CDI will garbage collect completed DataVolume. For VM uses DataVolume, virtink waits its DataVolume import to be completed before starting the VM. But with default CDI deployment, DataVolume will be GCed once completed. Hence virtink fails deploy VM with
build VM Pod: DataVolume.cdi.kubevirt.io "cdi-volume" not found
error.Perhaps it's worth mentioning this in CDI VM example. Tuning
.spec.config.dataVolumeTTLSeconds
ofcdi.kubevirt.io/v1beta1
works.(or adapting virtink to CDI's GC action?)