Claude on Question 13, Sep 10

The short answer: you can't kill a zombie — it's already dead

A zombie (state Z, also called a "defunct" process) isn't running. It has already exited. All that remains is an entry in the kernel's process table holding its exit status, waiting for its parent to call wait()/waitpid() and collect it. Since there's no code executing, signals have nowhere to go — SIGTERM, SIGKILL, anything you throw at it is simply ignored.

The real problem is the parent process, which is failing to reap its children.

Find the zombies and their parents

ps -eo pid,ppid,stat,cmd | awk '$3 ~ /^Z/'

Or for a specific zombie's parent:

ps -o ppid= -p <zombie_pid>

Step 1: Nudge the parent

Sending SIGCHLD tells the parent a child has terminated, which may prompt a well-behaved-but-sleeping parent to reap it:

kill -s SIGCHLD <parent_pid>

This often does nothing, because the usual cause is a parent with a missing or broken SIGCHLD handler. Worth a try since it's non-disruptive.

Step 2: Kill or restart the parent

When the parent dies, its orphaned children are re-parented to PID 1 (systemd or init), which reaps them immediately.

kill -TERM <parent_pid>      # try graceful first
kill -KILL <parent_pid>      # if it doesn't exit

If the parent is a managed service, restart it properly instead:

systemctl restart <service>

Step 3 (advanced): reap manually with gdb

If killing the parent is unacceptable — say it's a long-running database — attach a debugger and call waitpid on its behalf:

gdb -p <parent_pid>
(gdb) call waitpid(-1, 0, 1)   # 1 = WNOHANG
(gdb) detach
(gdb) quit

This briefly stops the process and pokes at its internals, so treat it as a last resort on anything you care about.

When to actually worry

A handful of zombies is harmless — they consume a PID slot and a few bytes of kernel memory, nothing more. They become a real problem only when they accumulate and exhaust the PID space:

cat /proc/sys/kernel/pid_max        # typically 32768 or 4194304
ps -eo stat | grep -c '^Z'          # current zombie count

If a process is leaking zombies steadily, that's a bug in the parent, and restarting it is the fix while you chase down the root cause.

A common gotcha: containers

Inside Docker/Kubernetes containers, PID 1 is often your application rather than a real init system, and most applications don't reap orphans. Zombies then pile up with no one to collect them. Fix it at the container level:

docker run --init your-image

Or bake in a minimal init like tini or dumb-init as your entrypoint. In Kubernetes, shareProcessNamespace: true also gives you a reaping pause container.

Only if PID 1 itself is wedged and refusing to reap — which is rare and indicates something badly wrong — is a reboot the answer.