Use Firecracker SDK in daemon
Replace the daemon's hand-rolled Firecracker process/socket client with the official firecracker-go-sdk while keeping the existing VM lifecycle and host-side disk and TAP setup intact. Build machine configs through the SDK, launch Firecracker through a sudo process runner, resolve the real VM PID after startup, and use the SDK client for Ctrl-Alt-Del instead of raw REST calls. Drop the unused cached Firecracker state and add focused adapter tests for config and process-runner wiring. Validated with go mod tidy, go test ./..., and make build. A live KVM/Firecracker smoke boot was not run in this environment.
This commit is contained in:
parent
ea72ea26fe
commit
2539800f5c
6 changed files with 1393 additions and 179 deletions
|
|
@ -4,13 +4,11 @@ import (
|
|||
"context"
|
||||
"errors"
|
||||
"fmt"
|
||||
"net"
|
||||
"os"
|
||||
"os/exec"
|
||||
"path/filepath"
|
||||
"strconv"
|
||||
"strings"
|
||||
"syscall"
|
||||
"time"
|
||||
|
||||
"banger/internal/api"
|
||||
|
|
@ -197,69 +195,32 @@ func (d *Daemon) startVMLocked(ctx context.Context, vm model.VMRecord, image mod
|
|||
if err != nil {
|
||||
return cleanupOnErr(err)
|
||||
}
|
||||
pid, err := d.startFirecrackerProcess(ctx, fcPath, apiSock, vm.Runtime.LogPath)
|
||||
machine, err := firecracker.NewMachine(ctx, firecracker.MachineConfig{
|
||||
BinaryPath: fcPath,
|
||||
VMID: vm.ID,
|
||||
SocketPath: apiSock,
|
||||
LogPath: vm.Runtime.LogPath,
|
||||
MetricsPath: vm.Runtime.MetricsPath,
|
||||
KernelImagePath: image.KernelPath,
|
||||
InitrdPath: image.InitrdPath,
|
||||
KernelArgs: system.BuildBootArgs(vm.Name, vm.Runtime.GuestIP, d.config.BridgeIP, d.config.DefaultDNS),
|
||||
RootDrivePath: vm.Runtime.DMDev,
|
||||
WorkDrivePath: vm.Runtime.WorkDiskPath,
|
||||
TapDevice: tap,
|
||||
VCPUCount: vm.Spec.VCPUCount,
|
||||
MemoryMiB: vm.Spec.MemoryMiB,
|
||||
})
|
||||
if err != nil {
|
||||
return cleanupOnErr(err)
|
||||
}
|
||||
vm.Runtime.PID = pid
|
||||
|
||||
if err := d.waitForSocket(ctx, apiSock); err != nil {
|
||||
if err := machine.Start(ctx); err != nil {
|
||||
vm.Runtime.PID = d.resolveFirecrackerPID(ctx, machine, apiSock)
|
||||
return cleanupOnErr(err)
|
||||
}
|
||||
if actualPID, err := d.findFirecrackerPID(ctx, apiSock); err == nil && actualPID > 0 {
|
||||
vm.Runtime.PID = actualPID
|
||||
}
|
||||
|
||||
client := firecracker.New(apiSock)
|
||||
if err := client.Put(ctx, "/machine-config", map[string]any{
|
||||
"vcpu_count": vm.Spec.VCPUCount,
|
||||
"mem_size_mib": vm.Spec.MemoryMiB,
|
||||
"smt": false,
|
||||
}); err != nil {
|
||||
vm.Runtime.PID = d.resolveFirecrackerPID(ctx, machine, apiSock)
|
||||
if err := d.ensureSocketAccess(ctx, apiSock); err != nil {
|
||||
return cleanupOnErr(err)
|
||||
}
|
||||
if err := client.Put(ctx, "/metrics", map[string]any{
|
||||
"metrics_path": vm.Runtime.MetricsPath,
|
||||
}); err != nil {
|
||||
return cleanupOnErr(err)
|
||||
}
|
||||
boot := map[string]any{
|
||||
"kernel_image_path": image.KernelPath,
|
||||
"boot_args": system.BuildBootArgs(vm.Name, vm.Runtime.GuestIP, d.config.BridgeIP, d.config.DefaultDNS),
|
||||
}
|
||||
if image.InitrdPath != "" {
|
||||
boot["initrd_path"] = image.InitrdPath
|
||||
}
|
||||
if err := client.Put(ctx, "/boot-source", boot); err != nil {
|
||||
return cleanupOnErr(err)
|
||||
}
|
||||
if err := client.Put(ctx, "/drives/rootfs", map[string]any{
|
||||
"drive_id": "rootfs",
|
||||
"path_on_host": vm.Runtime.DMDev,
|
||||
"is_root_device": true,
|
||||
"is_read_only": false,
|
||||
}); err != nil {
|
||||
return cleanupOnErr(err)
|
||||
}
|
||||
if err := client.Put(ctx, "/drives/work", map[string]any{
|
||||
"drive_id": "work",
|
||||
"path_on_host": vm.Runtime.WorkDiskPath,
|
||||
"is_root_device": false,
|
||||
"is_read_only": false,
|
||||
}); err != nil {
|
||||
return cleanupOnErr(err)
|
||||
}
|
||||
if err := client.Put(ctx, "/network-interfaces/eth0", map[string]any{
|
||||
"iface_id": "eth0",
|
||||
"host_dev_name": tap,
|
||||
}); err != nil {
|
||||
return cleanupOnErr(err)
|
||||
}
|
||||
if err := client.Put(ctx, "/actions", map[string]any{"action_type": "InstanceStart"}); err != nil {
|
||||
return cleanupOnErr(err)
|
||||
}
|
||||
fcConfig, _ := client.GetConfig(ctx)
|
||||
vm.Runtime.FirecrackerState = fcConfig
|
||||
if err := d.setDNS(ctx, vm.Name, vm.Runtime.GuestIP); err != nil {
|
||||
return cleanupOnErr(err)
|
||||
}
|
||||
|
|
@ -615,59 +576,6 @@ func (d *Daemon) firecrackerBinary() (string, error) {
|
|||
return path, nil
|
||||
}
|
||||
|
||||
func (d *Daemon) startFirecrackerProcess(ctx context.Context, fcBin, apiSock, logPath string) (int, error) {
|
||||
logFile, err := os.OpenFile(logPath, os.O_CREATE|os.O_WRONLY|os.O_APPEND, 0o644)
|
||||
if err != nil {
|
||||
return 0, err
|
||||
}
|
||||
cmd := exec.CommandContext(ctx, "sudo", "-n", fcBin, "--api-sock", apiSock)
|
||||
cmd.Stdout = logFile
|
||||
cmd.Stderr = logFile
|
||||
cmd.Stdin = nil
|
||||
cmd.SysProcAttr = &syscall.SysProcAttr{Setpgid: true}
|
||||
if err := cmd.Start(); err != nil {
|
||||
_ = logFile.Close()
|
||||
return 0, err
|
||||
}
|
||||
go func() {
|
||||
_ = cmd.Wait()
|
||||
_ = logFile.Close()
|
||||
}()
|
||||
return cmd.Process.Pid, nil
|
||||
}
|
||||
|
||||
func (d *Daemon) waitForSocket(ctx context.Context, apiSock string) error {
|
||||
deadline := time.Now().Add(15 * time.Second)
|
||||
var lastErr error
|
||||
for {
|
||||
if _, err := os.Stat(apiSock); err == nil {
|
||||
if err := d.ensureSocketAccess(ctx, apiSock); err != nil {
|
||||
lastErr = err
|
||||
} else {
|
||||
conn, dialErr := net.DialTimeout("unix", apiSock, 200*time.Millisecond)
|
||||
if dialErr == nil {
|
||||
_ = conn.Close()
|
||||
return nil
|
||||
}
|
||||
lastErr = dialErr
|
||||
}
|
||||
} else if !os.IsNotExist(err) {
|
||||
lastErr = err
|
||||
}
|
||||
if time.Now().After(deadline) {
|
||||
if lastErr != nil {
|
||||
return fmt.Errorf("firecracker api socket not ready: %s: %w", apiSock, lastErr)
|
||||
}
|
||||
return fmt.Errorf("firecracker api socket not ready: %s", apiSock)
|
||||
}
|
||||
select {
|
||||
case <-ctx.Done():
|
||||
return ctx.Err()
|
||||
case <-time.After(20 * time.Millisecond):
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
func (d *Daemon) ensureSocketAccess(ctx context.Context, apiSock string) error {
|
||||
if _, err := d.runner.RunSudo(ctx, "chown", fmt.Sprintf("%d:%d", os.Getuid(), os.Getgid()), apiSock); err != nil {
|
||||
return err
|
||||
|
|
@ -684,12 +592,24 @@ func (d *Daemon) findFirecrackerPID(ctx context.Context, apiSock string) (int, e
|
|||
return strconv.Atoi(strings.TrimSpace(string(out)))
|
||||
}
|
||||
|
||||
func (d *Daemon) resolveFirecrackerPID(ctx context.Context, machine *firecracker.Machine, apiSock string) int {
|
||||
if pid, err := d.findFirecrackerPID(ctx, apiSock); err == nil && pid > 0 {
|
||||
return pid
|
||||
}
|
||||
if machine != nil {
|
||||
if pid, err := machine.PID(); err == nil && pid > 0 {
|
||||
return pid
|
||||
}
|
||||
}
|
||||
return 0
|
||||
}
|
||||
|
||||
func (d *Daemon) sendCtrlAltDel(ctx context.Context, vm model.VMRecord) error {
|
||||
if err := d.ensureSocketAccess(ctx, vm.Runtime.APISockPath); err != nil {
|
||||
return err
|
||||
}
|
||||
client := firecracker.New(vm.Runtime.APISockPath)
|
||||
return client.Put(ctx, "/actions", map[string]any{"action_type": "SendCtrlAltDel"})
|
||||
return client.SendCtrlAltDel(ctx)
|
||||
}
|
||||
|
||||
func (d *Daemon) waitForExit(ctx context.Context, pid int, apiSock string, timeout time.Duration) error {
|
||||
|
|
@ -748,7 +668,6 @@ func clearRuntimeHandles(vm *model.VMRecord) {
|
|||
vm.Runtime.COWLoop = ""
|
||||
vm.Runtime.DMName = ""
|
||||
vm.Runtime.DMDev = ""
|
||||
vm.Runtime.FirecrackerState = nil
|
||||
}
|
||||
|
||||
func (d *Daemon) setDNS(ctx context.Context, vmName, guestIP string) error {
|
||||
|
|
@ -794,8 +713,6 @@ func (d *Daemon) requireStartPrereqs(ctx context.Context) error {
|
|||
ctx,
|
||||
"sudo",
|
||||
"ip",
|
||||
"curl",
|
||||
"jq",
|
||||
"dmsetup",
|
||||
"losetup",
|
||||
"blockdev",
|
||||
|
|
|
|||
Loading…
Add table
Add a link
Reference in a new issue