Add image metadata, allow flags to be overridden, and tzdata (#12)

That last one was a major omission and would cause the daemon to hang
when run in a container.

Dockerfile

@@ -1,4 +1,16 @@
 FROM alpine:latest
+
+LABEL org.opencontainers.image.title=NXTP
+LABEL org.opencontainers.image.description="An NXTP client and server"
+LABEL org.opencontainers.image.vendor="Keith Gaughan"
+LABEL org.opencontainers.image.licenses=MIT
+LABEL org.opencontainers.image.url=https://github.com/kgaughan/nxtp
+LABEL org.opencontainers.image.source=https://github.com/kgaughan/nxtp
+LABEL org.opencontainers.image.documentation=https://kgaughan.github.io/nxtp/
+
+RUN apk --no-cache add tzdata
 COPY nxtp .
+USER nobody
 EXPOSE 12300
-ENTRYPOINT ["/nxtp", "-endpoint", ":12300"]
+ENTRYPOINT ["/nxtp"]
+CMD ["-endpoint", "0.0.0.0:12300"]

README.md

@@ -57,6 +57,14 @@ Restart = on-failure
 Under `contrib` is a NetBSD rc script. It expects _daemond_ to be installed
 to help daemonise the server, which you can do with `pkgin install daemond`.
 
+## Running under Docker
+
+To start the daemon in a Docker container, run:
+
+```sh
+docker run -d -p 12300:12300 ghcr.io/kgaughan/nxtp:latest
+```
+
 ## Building
 
 Run the following:

contrib/nxtp.service

@@ -0,0 +1,12 @@
+[Unit]
+Description = Network neXt Time Protocol (NXTP) server
+After = network.target
+
+[Service]
+Type = simple
+ExecStart = /usr/bin/nxtp
+KillMode = process
+RestartSec = 5s
+Restart = on-failure
+
+# vi:set ft=dosini: