unarr/internal/library/mediainfo/harden_linux_test.go
Deivid Soto c82826bf68
Some checks failed
CI / Test (push) Failing after 6m21s
CI / Build (push) Successful in 1m34s
CI / Build-1 (push) Successful in 2m0s
CI / Build-2 (push) Successful in 1m33s
CI / Build-3 (push) Successful in 1m38s
CI / Build-4 (push) Successful in 1m35s
CI / Build-5 (push) Successful in 1m38s
CI / Lint (push) Failing after 2m34s
CI / Coverage (push) Failing after 2m44s
CI / Vet (push) Successful in 2m3s
fix(trickplay): stop scan-time sprite generation from saturating the host
Trickplay sprite generation (one full-decode ffmpeg pass per file) could pin a
machine: multiple agents on the same library decoded the same 4K file at once, no
CPU throttling, and crashed/restarted agents orphaned ffmpeg to init (it ran the
full 45-min decode to completion). Stacked orphans spiked a box to load ~140.

- Single-flight lock: O_CREATE|O_EXCL .lock in the shared sidecar dir so two
  agents watching the same library never decode the same file twice (stale locks
  reclaimed after a TTL). Returns ErrTrickplayInProgress → prewarm skips, not fail.
- Load gate: defer the heavy decode until 1-min load ≤ max(ratio×NumCPU, 1.5),
  capped at 15 min so it throttles without ever becoming a permanent off-switch on
  busy / small hosts. New knob library.prewarm_max_load_ratio (default 0.7).
- Concurrency: trickSem caps trickplay to ONE decode at a time per agent.
- CPU priority: setLowCPUPriority (nice 19) alongside the existing idle ionice.
- No orphans: hardenCmd sets Setpgid + Pdeathsig=SIGKILL, with runtime.LockOSThread
  around the child so the kernel kills ffmpeg exactly when the agent dies (and not
  spuriously — golang/go#27505).

Tests: single-flight/stale-reclaim, load-gate immediate/cancel, and an e2e
Pdeathsig orphan-kill check.
2026-06-04 08:25:00 +02:00

65 lines
1.9 KiB
Go

//go:build linux
package mediainfo
import (
"fmt"
"os"
"os/exec"
"strconv"
"strings"
"syscall"
"testing"
"time"
)
// TestHardenCmd_KillsChildOnParentDeath is the e2e guarantee for the orphan fix:
// a child spawned with hardenCmd must be SIGKILL'd by the kernel the instant its
// parent process dies (Pdeathsig), so an agent crash/restart can never leave an
// ffmpeg running to ppid 1. It re-execs this test binary as a short-lived helper
// that starts `sleep`, prints the sleep PID, then exits — and asserts that PID is
// gone afterwards.
func TestHardenCmd_KillsChildOnParentDeath(t *testing.T) {
if os.Getenv("UNARR_PDEATHSIG_CHILD") == "1" {
// Helper role: start a hardened long sleep, announce its PID, then exit so
// the kernel fires Pdeathsig on it.
cmd := exec.Command("sleep", "120")
hardenCmd(cmd)
if err := cmd.Start(); err != nil {
fmt.Println("ERR", err)
os.Exit(2)
}
fmt.Println(cmd.Process.Pid)
os.Exit(0)
}
helper := exec.Command(os.Args[0], "-test.run=TestHardenCmd_KillsChildOnParentDeath", "-test.v")
helper.Env = append(os.Environ(), "UNARR_PDEATHSIG_CHILD=1")
out, err := helper.Output()
if err != nil {
t.Fatalf("helper run: %v (out=%q)", err, out)
}
var sleepPID int
for _, line := range strings.Split(string(out), "\n") {
if n, perr := strconv.Atoi(strings.TrimSpace(line)); perr == nil && n > 0 {
sleepPID = n
break
}
}
if sleepPID == 0 {
t.Fatalf("could not parse child PID from helper output: %q", out)
}
// Give the kernel a moment to deliver SIGKILL after the helper exited.
deadline := time.Now().Add(3 * time.Second)
for time.Now().Before(deadline) {
if syscall.Kill(sleepPID, 0) != nil {
return // process gone → Pdeathsig worked
}
time.Sleep(50 * time.Millisecond)
}
// Cleanup if it somehow survived, then fail.
_ = syscall.Kill(sleepPID, syscall.SIGKILL)
t.Fatalf("child %d survived parent death — Pdeathsig not applied (orphan leak)", sleepPID)
}