Restore backup container leaves zombie process

Nobody has claimed this yet.

After implementing with docker compose as specified in the documentation, a zombie process is always left behind after running `docker compose up -d`. I believe this to be caused by the `restore-backup` service that immediately exits.

Here is my `compose.yml`:
```yml
services:
# TerraFirmaCraft server
mc:
image: itzg/minecraft-server
ports:
- "25565:25565"
environment:
UID: 1002
MEMORY: 6G
EULA: "TRUE"

# server settings
LEVEL_TYPE: tfc:overworld
SPAWN_PROTECTION: 0
ICON: /icon.png
DIFFICULTY: normal
MAX_PLAYERS: 5
ENABLE_WHITELIST: true

# mod settings
MOD_PLATFORM: AUTO_CURSEFORGE
CF_API_KEY: ${CF_API_KEY}
CF_SLUG: "terrafirmacraft-worlds-apart"
CF_FORCE_SYNCHRONIZE: true

depends_on:
restore-backup:
condition: service_completed_successfully
volumes:
- ./server-data/tfc/data:/data
- ./server-data/tfc/icon.png:/icon.png:ro
- ./server-data/tfc/datapacks:/datapacks:ro
# Backup restore service
restore-backup:
image: itzg/mc-backup
restart: "no"
entrypoint: restore-tar-backup
volumes:
- ./server-data/tfc/data:/data
- ./server-data/tfc/backups:/backups:ro
# Backup service
backups:
image: itzg/mc-backup
depends_on:
mc:
condition: service_healthy
environment:
BACKUP_INTERVAL: "8h"
RCON_HOST: mc
INITIAL_DELAY: 0
volumes:
- ./server-data/tfc/data:/data:ro
- ./server-data/tfc/backups:/backups
```

Start with the compose.yml configuration in the issue, especially the restore-backup service and its restore-tar-backup entrypoint, then reproduce docker compose up -d. Determine why the immediately exiting restore-backup service leaves a zombie process; done means the restore completes without leaving that process behind.

Written by the indexing model from the issue text.