Since a few months it is now possible to build an upstream QEMU which includes only our supported VCPU accelerator.
This simple project is about experimenting with replacing our usual symlink qemu-kvm with an actual binary, looking at the impact on the end-to-end setup, including libvirt, virt-manager (maybe kubevirt too?)
Goal for this Hackweek
Build the new qemu-kvm QEMU, fix issues as they present themselves for fitting into the stack, try to live migrate a few VMs, learn whether this could be an interesting direction/option to provide in the future in addition to qemu-system-xxx as our supported binary.
For code changes to QEMU we could use f.e.:
access to some identical machines for live migration, help with problem solving on the stack would be helpful.
Looking for hackers with the skills:
Nothing? Add some keywords!
This project is part of:
Hack Week 20
This project is one of its kind!