clonetool/manager.go

package main

import (
	"bufio"
	"errors"
	"fmt"
	"os"
	"strings"
	"time"
)

const defaultBlockSize = 4 * 1024 * 1024
const defaultConnectTimeoutSec = 8

// SyncConfig holds everything the `sync` command needs. It is also handed
// (the transport-relevant fields of it) to control agents as part of
// connect_push/connect_pull requests, so they know how to reach the peer.
type SyncConfig struct {
	Job               string
	Source            string
	Dest              string
	BlockSize         int64
	Yes               bool
	Sudo              string // auto | always | never
	Deploy            bool   // copy this binary to remote hosts that lack it
	ConnectTimeoutSec int
	SSHBin            string
	SSHOpts           []string
	RemoteBin         string
	ManagerHost       string
}

func runSync(cfg SyncConfig) error {
	srcSpec, err := parseSpec(cfg.Source)
	if err != nil {
		return fmt.Errorf("--source: %w", err)
	}
	dstSpec, err := parseSpec(cfg.Dest)
	if err != nil {
		return fmt.Errorf("--dest: %w", err)
	}
	if err := checkNotSame(srcSpec, dstSpec); err != nil {
		return err
	}

	// Make sure each remote endpoint has a runnable clonetool, copying this
	// binary over if not (unless --deploy=false). A host that appears on
	// both sides is only probed once.
	srcRemoteBin, dstRemoteBin := cfg.RemoteBin, cfg.RemoteBin
	if !srcSpec.IsLocal() {
		if srcRemoteBin, err = resolveRemoteBin(&cfg, srcSpec, "source"); err != nil {
			return err
		}
	}
	if !dstSpec.IsLocal() {
		if !srcSpec.IsLocal() && sameHost(srcSpec, dstSpec) {
			dstRemoteBin = srcRemoteBin
		} else if dstRemoteBin, err = resolveRemoteBin(&cfg, dstSpec, "dest"); err != nil {
			return err
		}
	}

	srcCtrl, srcInfo, srcSudo, err := bringUpController(srcSpec, "source", &cfg, srcRemoteBin, srcSpec.Path)
	if err != nil {
		return fmt.Errorf("source: %w", err)
	}
	defer srcCtrl.Close()
	if !srcInfo.Exists {
		return fmt.Errorf("source %s does not exist", srcSpec)
	}

	dstCtrl, dstInfo, dstSudo, err := bringUpController(dstSpec, "dest", &cfg, dstRemoteBin, dstSpec.Path)
	if err != nil {
		return fmt.Errorf("dest: %w", err)
	}
	defer dstCtrl.Close()

	targetSize, err := computeTargetSize(srcSpec, srcInfo, dstSpec, dstInfo, cfg.Yes)
	if err != nil {
		return err
	}

	if err := dstCtrl.Prepare(dstSpec.Path, targetSize); err != nil {
		return fmt.Errorf("prepare destination: %w", err)
	}

	blockCount := (targetSize + cfg.BlockSize - 1) / cfg.BlockSize

	pp := newProgressPrinter(cfg.BlockSize)
	cb := transferCallbacks{onProgress: pp.print}

	bothLocal := srcSpec.IsLocal() && dstSpec.IsLocal()
	srcHost, srcUser := resolveConnectHost(srcSpec, &cfg)
	dstHost, dstUser := resolveConnectHost(dstSpec, &cfg)

	pushReq := CtrlMsg{
		Path: srcSpec.Path, Size: targetSize, BlockSize: cfg.BlockSize,
		PeerHost: dstHost, PeerUser: dstUser, PeerPath: dstSpec.Path, PeerLocal: bothLocal,
		RemoteBin: dstRemoteBin, SSHBin: cfg.SSHBin, SSHOpts: cfg.SSHOpts, ConnectTimeoutSec: cfg.ConnectTimeoutSec,
		Sudo: dstSudo,
	}
	if bothLocal {
		fmt.Fprintf(os.Stderr, "both source and dest are local; syncing directly (no ssh) ...\n")
	} else {
		fmt.Fprintf(os.Stderr, "attempting push %s -> %s ...\n", srcSpec, dstSpec)
	}
	ok, reason, err := srcCtrl.ConnectPush(pushReq, cb)
	if err != nil {
		return err
	}
	if !ok {
		pp.finish()
		fmt.Fprintf(os.Stderr, "push not possible (%s); trying pull %s <- %s ...\n", reason, dstSpec, srcSpec)
		pullReq := CtrlMsg{
			Path: dstSpec.Path, Size: targetSize, BlockSize: cfg.BlockSize,
			PeerHost: srcHost, PeerUser: srcUser, PeerPath: srcSpec.Path, PeerLocal: bothLocal,
			RemoteBin: srcRemoteBin, SSHBin: cfg.SSHBin, SSHOpts: cfg.SSHOpts, ConnectTimeoutSec: cfg.ConnectTimeoutSec,
			Sudo: srcSudo,
		}
		ok2, reason2, err2 := dstCtrl.ConnectPull(pullReq, cb)
		if err2 != nil {
			return err2
		}
		if !ok2 {
			return fmt.Errorf(
				"could not establish a direct connection in either direction (push: %s; pull: %s); "+
					"run the manager on the source or destination host, or set up SSH connectivity in at least one direction",
				reason, reason2)
		}
	}

	pp.finish()
	label := cfg.Job
	if label == "" {
		label = fmt.Sprintf("%s -> %s", srcSpec, dstSpec)
	}
	fmt.Fprintf(os.Stderr, "done: %s, %d blocks\n", label, blockCount)
	return nil
}

// bringUpController starts a control agent for spec and does its initial
// stat. With --sudo=always the agent is elevated from the start; with
// --sudo=auto a permission error on the stat triggers one transparent
// restart under sudo. The returned bool reports whether the agent (and any
// peer helper it later spawns for this side) is running elevated.
func bringUpController(spec Spec, tag string, cfg *SyncConfig, remoteBin, probePath string) (*Controller, PathInfo, bool, error) {
	sudo := cfg.Sudo == "always"
	c, err := startController(spec, tag, cfg, remoteBin, sudo)
	if err != nil {
		return nil, PathInfo{}, sudo, err
	}
	info, err := c.Stat(probePath)
	if err == nil {
		return c, info, sudo, nil
	}
	if errors.Is(err, errNeedPriv) && cfg.Sudo == "auto" && !sudo {
		fmt.Fprintf(os.Stderr, "%s: permission denied on %s; retrying via sudo ...\n", tag, probePath)
		c.Close()
		sudo = true
		if c, err = startController(spec, tag, cfg, remoteBin, true); err != nil {
			return nil, PathInfo{}, sudo, err
		}
		if info, err = c.Stat(probePath); err == nil {
			return c, info, sudo, nil
		}
	}
	c.Close()
	if sudo {
		return nil, PathInfo{}, sudo, fmt.Errorf(
			"%w; if this is a sudo failure, configure passwordless sudo for clonetool on %s or run the manager as root",
			err, hostLabel(spec))
	}
	return nil, PathInfo{}, sudo, err
}

func sameHost(a, b Spec) bool {
	return strings.EqualFold(a.Host, b.Host) && a.User == b.User
}

func hostLabel(spec Spec) string {
	if spec.IsLocal() {
		return "this machine"
	}
	return spec.Host
}

func resolveConnectHost(spec Spec, cfg *SyncConfig) (host, user string) {
	if !spec.IsLocal() {
		return spec.Host, spec.User
	}
	host = cfg.ManagerHost
	if host == "" {
		if h, err := os.Hostname(); err == nil {
			host = h
		}
	}
	return host, ""
}

// computeTargetSize applies the sizing/truncation rules: a block-device
// destination can't grow, so the source must fit inside it (sync exactly
// min(src,dst), leaving the remainder of the device untouched); a
// regular-file destination is truncated to the source's size, shrinking
// with confirmation if it currently holds more data than that.
func computeTargetSize(srcSpec Spec, srcInfo PathInfo, dstSpec Spec, dstInfo PathInfo, yes bool) (int64, error) {
	if dstInfo.Exists && dstInfo.IsDevice {
		if srcInfo.Size > dstInfo.Size {
			return 0, fmt.Errorf("source %s (%s) is larger than destination device %s (%s); a device can't be grown",
				srcSpec, humanBytes(srcInfo.Size), dstSpec, humanBytes(dstInfo.Size))
		}
		return srcInfo.Size, nil
	}
	if dstInfo.Exists && dstInfo.Size > srcInfo.Size {
		if !yes {
			if !confirmShrink(dstSpec, dstInfo.Size, srcInfo.Size) {
				return 0, fmt.Errorf("aborted: destination %s would be truncated from %s to %s", dstSpec, humanBytes(dstInfo.Size), humanBytes(srcInfo.Size))
			}
		}
	}
	return srcInfo.Size, nil
}

func confirmShrink(dstSpec Spec, oldSize, newSize int64) bool {
	fmt.Fprintf(os.Stderr, "destination %s is %s; it will be truncated to %s to match the source. Continue? [y/N] ",
		dstSpec, humanBytes(oldSize), humanBytes(newSize))
	reader := bufio.NewReader(os.Stdin)
	line, _ := reader.ReadString('\n')
	line = strings.TrimSpace(strings.ToLower(line))
	return line == "y" || line == "yes"
}

// progressPrinter renders a single rewriting status line on stderr for both
// phases of a run: "scan" while a side hashes the destination, then the
// block transfer itself (with a rolling copy rate).
type progressPrinter struct {
	blockSize  int64
	start      time.Time
	lastT      time.Time
	lastCopied int64
	rate       float64 // copied blocks/sec, smoothed
	active     bool
}

func newProgressPrinter(blockSize int64) *progressPrinter {
	now := time.Now()
	return &progressPrinter{blockSize: blockSize, start: now, lastT: now}
}

func (p *progressPrinter) print(m CtrlMsg) {
	p.active = true
	if m.Phase == "scan" {
		fmt.Fprintf(os.Stderr, "\r  scanning destination %s %s/%s blocks          ",
			progressBar(m.Copied, m.TotalBlocks),
			formatCount(m.Copied), formatCount(m.TotalBlocks))
		return
	}

	done := m.Copied + m.Skipped
	total := m.TotalBlocks
	if total <= 0 {
		total = done
	}

	now := time.Now()
	if dt := now.Sub(p.lastT).Seconds(); dt >= 0.4 {
		inst := float64(m.Copied-p.lastCopied) / dt
		if p.rate == 0 {
			p.rate = inst
		} else {
			p.rate = 0.6*p.rate + 0.4*inst
		}
		p.lastT, p.lastCopied = now, m.Copied
	}
	mib := p.rate * float64(p.blockSize) / (1024 * 1024)

	fmt.Fprintf(os.Stderr, "\r  syncing %s %s/%s  copied=%s skipped=%s  %6.1f MiB/s          ",
		progressBar(done, total), formatCount(done), formatCount(total),
		formatCount(m.Copied), formatCount(m.Skipped), mib)
}

// finish ends the current status line with a newline so following output
// (and the shell prompt) starts clean.
func (p *progressPrinter) finish() {
	if p.active {
		fmt.Fprintln(os.Stderr)
		p.active = false
	}
}

func progressBar(done, total int64) string {
	const w = 22
	pct := 0
	if total > 0 {
		pct = int(done * 100 / total)
	}
	if pct > 100 {
		pct = 100
	}
	n := pct * w / 100
	return fmt.Sprintf("[%s%s] %3d%%", strings.Repeat("=", n), strings.Repeat(" ", w-n), pct)
}

func formatCount(n int64) string {
	switch {
	case n >= 1_000_000:
		return fmt.Sprintf("%.1fM", float64(n)/1e6)
	case n >= 10_000:
		return fmt.Sprintf("%.0fk", float64(n)/1e3)
	default:
		return fmt.Sprintf("%d", n)
	}
}