Remove stale pid and socket files

[MorganRodgers]

Fixes #11.

[ericfranz]

@treydock is there any reason why always removing stale pid files when calling nginx_stage nginx_clean would be problematic?

Is there any utility in retaining stale pid files? I could imagine there would be a case where it would help with diagnosing an issue, but not sure. I'm not sure that nginx_stage nginx_clean prints anything to stdout or stderr or if we started printing to stderr a message when removing stale pid files if that would be captured by anything.

The current approach was to not change behavior and add a new flag that removed pid files, but then again user expectation was different:

I would have expected nginx_stage nginx_clean to remove those stale pidfiles

[treydock]

@ericfranz I think removing stale PIDs is fine. I'd suggest writing the removed PID file and the actual PID (if available) to stdout or stderr and then we would update the deployed cron job to look like the following:

/opt/ood/nginx_stage/sbin/nginx_stage nginx_clean 2>&1 | logger -t nginx_stage

This would be the simplest approach for nginx_stage and would still write the output to syslog and not email people the output. I like this because then it's very easy to setup Splunk or other tools to tell me when something pops up in the logs.

[ericfranz]

@MorganRodgers lets change nginx clean to always remove stale pid files and just output the stale pid removed to stdout stale pid file removed: /var/run/nginx/kilian/passenger.pid

The code should be a whole lot simpler then!

[ericfranz]

See comment