GPT on Question 14, Sep 20

A process cannot clean up children after it has died. In Bash, the usual approach is to install an EXIT trap and terminate known child PIDs before exiting:

#!/usr/bin/env bash

children=()

cleanup() {
    rc=$?
    trap - EXIT HUP INT TERM

    if ((${#children[@]})); then
        kill -TERM -- "${children[@]}" 2>/dev/null || :
        wait "${children[@]}" 2>/dev/null || :
    fi

    exit "$rc"
}

trap cleanup EXIT
trap 'exit 129' HUP
trap 'exit 130' INT
trap 'exit 143' TERM

worker1 &
children+=("$!")

worker2 &
children+=("$!")

wait "${children[@]}"
children=()

Important limitations:

• •Traps do not run for SIGKILL, kernel crashes, power loss, etc.
• •This handles only the PIDs you track, not arbitrary grandchildren.
• •pkill -P $$ only finds direct children and stops working once they are reparented.
• •Avoid kill 0 unless the script is deliberately placed in its own process group; otherwise it may kill its caller too.

For robust cleanup of an entire process tree, use an external supervisor or cgroup, such as a systemd service with:

[Service]
ExecStart=/path/to/script
KillMode=control-group

On Linux, another option for individual children is a small wrapper using prctl(PR_SET_PDEATHSIG, SIGTERM), causing the kernel to signal the child when its parent dies. Bash has no built-in interface for that, and each relevant process must arrange it itself.

In short: use an EXIT trap for normal Bash cleanup; use systemd/cgroups when cleanup must survive the parent being forcibly killed.