Re: Parallel input from growing file?

[Hans Schou]

Start reading this chapter in the man page:

  EXAMPLE: GNU Parallel as queue system/batch manager

You are not the first who (mis-)use Gnu Parallel as a queue manager.

/hans

2014-12-14 12:01 GMT+01:00 xmoon 2000 <xmoon2000@googlemail.com>:

Hi,

Currently I read jobs from a file on my c drive that gets written to
by other scripts, and my master script runs each job found in
background with &. I want to use parallel instead of my own script.

I would like parallel to monitor this file and run any jobs in it with
as many cores as are available, or wait until new jobs are appended to
the file. Running 24/7.

What is the syntax for doing this?

My jobFile looks like this:


/myApp prm1 parm2 parm2
/myApp parm1 parm2
/myApp parm1 parm2 parm3 parm4

I could do:

tail -f jobFile  | parallel

Is this the best approach?