dockerd: adjust shutdown timeout to allow graceful shutdown of containers

Maintainer: @G-M0N3Y-2503

By default procd has a shutdown timeout of 5 seconds, procd initially sends a SIGTERM kill signal to allow docker service to shutdown gracefully, after waiting 5 seconds, if the service is still running, it will send a SIGKILL signal forcing the service is shutdown forcefully and immediately without consideration even if the container is still writing a a file to the disk.

This behavior often causes severe data corruption and partial writes, especially for Database services like mysql, postgres, influxdb, prometheus (tsdb), SQLite-based services, etc. This becomes a major issue especially when you're frequently restarting your device/computer (x86).

I'd like to propose increasing the term_timeout in procd for docker service to at least 5 minutes AND allowing the configuration of shutdown-timeout in daemon.json through UCI.

Proposed changes in /etc/init.d/dockerd

in start_service()

procd_set_param term_timeout 300  # e.g. 5 minutes to wait for shutdown

in process_config()

...
config_get shutdown_timeout globals shutdown_timeout ""
...
json_add_int "shutdown-timeout" "${shutdown_timeout}"

Start with /etc/init.d/dockerd, especially start_service() and process_config(), and read how procd service parameters and UCI globals are handled there. Implement the requested 300-second term_timeout and daemon.json shutdown-timeout configuration, then verify the generated Docker configuration and service shutdown behavior.