How to find the .pid file for a given process

Solution 1:

You'll usually find the PID files for daemonized processes in /var/run/ on Redhat/CentOS-style systems.

Short of that, you can always look in the process init script. For instance, the SSH daemon is started with the script in /etc/init.d/sshd. Sometimes the PID will be defined there (search for pid, PID, PIDFILE, PID_FILE, etc.).

However, most other daemons on RHEL-style systems source the /etc/init.d/functions script for some common features.

# Set $pid to pids from /var/run* for {program}.  $pid should be declared
# local in the caller.
# Returns LSB exit code for the 'status' action.
__pids_var_run() {
        local base=${1##*/}
        local pid_file=${2:-/var/run/$base.pid}

For anything that sources /etc/init.d/functions, the PID will live in /var/run/*.pid.

For custom applications, the PID will be defined in a wrapper script (hopefully). Most developers I know follow the same convention as the daemons above, though.

If you do encounter something without a PID file, remember that Monit can monitor on a process string patern as well.

Solution 2:

Another approach I took:

I have a database server running in embedded mode, and the data are within the containing application's directory.

The database has something like a .pid file, but it calls it lock file. To locate this lock file, I listed all files held open by the app:

$ ls -l /proc/18264/fd | cut -d'>' -f2

That gave me a long list including sockets, pipes, server files etc. Few filters and I got to what I needed:

$ ls -l /proc/18264/fd | cut -d'>' -f2 | grep /home/ | cut -b40- | sort | uniq | grep titan

/windup/reports/group_report.LJfZVIavURqg.report/graph/titangraph/00000000.jdb
/windup/reports/group_report.LJfZVIavURqg.report/graph/titangraph/je.info.0
/windup/reports/group_report.LJfZVIavURqg.report/graph/titangraph/je.info.0.lck
/windup/reports/group_report.LJfZVIavURqg.report/graph/titangraph/je.lck