Single-node k8s on qemu

[yellowhat]

Hi,

I am familiarizing with talos by running a single-node k8s cluster using qemu.

I am able to use it by using vmlinuz-amd64 and initramfs-amd64.xz:

Install to VM

qemu-img create -f qcow2 os.qcow2 20G

qemu-system-x86_64 \
    -name VM \
    -enable-kvm \
    -machine "type=q35,accel=kvm" \
    -cpu host \
    -bios /usr/share/ovmf/OVMF.fd \
    -smp "cores=2,threads=2,sockets=1" \
    -m 8G \
    -nic "user,model=virtio-net-pci,net=172.16.0.0/24,hostfwd=tcp::6443-:6443,hostfwd=tcp::50000-:50000" \
    -drive "file=os.qcow2,if=virtio,format=qcow2" \
    -kernel vmlinuz \
    -initrd initramfs.xz \
    -append "talos.platform=metal slab_nomerge pti=on" \
    -display vnc=0.0.0.0:0 \
    -daemonize \
    -pidfile /tmp/vm.pid

talosctl gen config my https://172.16.0.15:6443/ \
    --install-disk /dev/vda \
    --additional-sans 127.0.0.1 \
    --with-examples=false \
    --with-docs=false

talosctl config endpoint 127.0.0.1
talosctl config node 127.0.0.1

# Wait to boot
until talosctl get disks --insecure --nodes 127.0.0.1; do sleep 1; done

# Install to disk
talosctl apply-config --file controlplane.yaml --insecure --nodes 127.0.0.1

# Wait to reboot
until talosctl get disks; do sleep 1; done

talosctl bootstrap

As talos provides images, I would like to use them.

I am using a raw generated with:

---
customization:
  systemExtensions:
    officialExtensions:
      - siderolabs/qemu-guest-agent

From raw image

talosctl gen config my https://172.16.0.15:6443/ \
    --install-disk /dev/vda \
    --additional-sans 127.0.0.1 \
    --with-examples=false \
    --with-docs=false

cloud-localds vm_config.raw controlplane.yaml --verbose

qemu-system-x86_64 \
    -name VM \
    -enable-kvm \
    -machine "type=q35,accel=kvm" \
    -cpu host \
    -bios /usr/share/ovmf/OVMF.fd \
    -smp "cores=2,threads=2,sockets=1" \
    -m 8G \
    -nic "user,model=virtio-net-pci,net=172.16.0.0/24,hostfwd=tcp::6443-:6443,hostfwd=tcp::50000-:50000" \
    -drive "file=img.raw,if=virtio,format=raw" \
    -drive "file=vm_config.raw,if=virtio,format=raw" \
    -display vnc=0.0.0.0:0 \
    -daemonize \
    -pidfile /tmp/vm.pid

talosctl config endpoint 127.0.0.1
talosctl config node 127.0.0.1

# Wait for boot
until talosctl get disks; do sleep 1; done

talosctl get members

But it does not bootstrap the cluster:

$ talosctl bootstrap
error executing bootstrap: 1 error occurred:
	* 127.0.0.1: rpc error: code = FailedPrecondition desc = bootstrap is not available yet

$ talosctl dmesg
127.0.0.1: user: warning: [2025-04-30T11:45:53.795743293Z]: [talos] node identity established {"component": "controller-runtime", "controller": "cluster.NodeIdentityController", "node_id": "BJGF7iTog63lnBIe7cEWh3zz3D0dxgGQJ4wRPkjOG2L"}
127.0.0.1: user: warning: [2025-04-30T11:45:53.796537293Z]: [talos] setting hostname {"component": "controller-runtime", "controller": "network.HostnameSpecController", "hostname": "talos-hpt-e2f", "domainname": ""}
127.0.0.1: user: warning: [2025-04-30T11:45:53.796589293Z]: [talos] setting hostname {"component": "controller-runtime", "controller": "network.HostnameSpecController", "hostname": "talos-hpt-e2f", "domainname": ""}
127.0.0.1: user: warning: [2025-04-30T11:45:53.796602293Z]: [talos] phase memorySizeCheck (4/15): 1 tasks(s)
127.0.0.1: user: warning: [2025-04-30T11:45:53.796686293Z]: [talos] task memorySizeCheck (1/1): starting
127.0.0.1: user: warning: [2025-04-30T11:45:53.804110293Z]: [talos] task memorySizeCheck (1/1): memory size is OK
127.0.0.1: user: warning: [2025-04-30T11:45:53.804115293Z]: [talos] task memorySizeCheck (1/1): memory size is 7917 MiB
127.0.0.1: user: warning: [2025-04-30T11:45:53.804128293Z]: [talos] task memorySizeCheck (1/1): done, 7.439981ms
127.0.0.1: user: warning: [2025-04-30T11:45:53.804141293Z]: [talos] phase memorySizeCheck (4/15): done, 7.544607ms
127.0.0.1: user: warning: [2025-04-30T11:45:53.804145293Z]: [talos] phase diskSizeCheck (5/15): 1 tasks(s)
127.0.0.1: user: warning: [2025-04-30T11:45:53.804166293Z]: [talos] task diskSizeCheck (1/1): starting
127.0.0.1: user: warning: [2025-04-30T11:45:53.804229293Z]: [talos] service[apid](Preparing): Running pre state
127.0.0.1: user: warning: [2025-04-30T11:45:53.804395293Z]: [talos] service[apid](Preparing): Creating service runner
127.0.0.1: user: warning: [2025-04-30T11:45:53.876904293Z]: [talos] volume status {"component": "controller-runtime", "controller": "block.VolumeManagerController", "volume": "EPHEMERAL", "phase": "failed -> failed", "error": "no disks matched for volume"}
127.0.0.1: user: warning: [2025-04-30T11:45:53.932103293Z]: [talos] service[apid](Running): Started task apid (PID 2068) for container apid
127.0.0.1: user: warning: [2025-04-30T11:45:54.807276293Z]: [talos] service[apid](Running): Health check successful
127.0.0.1: user: warning: [2025-04-30T11:45:58.949729293Z]: [talos] bootstrap request received
127.0.0.1: user: warning: [2025-04-30T11:46:12.939018293Z]: [talos] bootstrap request received
127.0.0.1: user: warning: [2025-04-30T11:46:13.112158293Z]: [talos] kubernetes endpoint watch error {"component": "controller-runtime", "controller": "k8s.EndpointController", "error": "failed to list *v1.Endpoints: Get \"https://172.16.0.15:6443/api/v1/namespaces/default/endpoints?fieldSelector=metadata.name%3Dkubernetes&limit=500&resourceVersion=0\": dial tcp 172.16.0.15:6443: connect: connection refused"}
127.0.0.1: user: warning: [2025-04-30T11:46:20.721004293Z]: [talos] volume status {"component": "controller-runtime", "controller": "block.VolumeManagerController", "volume": "EPHEMERAL", "phase": "failed -> failed", "error": "no disks matched for volume"}

Any suggestions?

[smira]

99% percent is that your disk is too small to allocate EPHEMERAL on it.

[yellowhat]

I just discovered at the sametime.

I was missing: qemu-img resize img.raw 20G

Anyway thanks for the prompt response