File-based job distribution for everyone
File-based distribution of jobs to CPUs on Unix-PCs.
There are powerful tools for automatic job distribution, but for smaller use cases, installation and maintenance effort is too high. In fjd, the hurdle to install and use should be very low (assigning jobs works via files in your home directory).
Jobs can be written in any language, fjd is only there to assign jobs to worker threads. Pull-model: idle workers get their next job from a queue. Works under the assumption that all CPUs are in a local network and can access a shared home directory.
Start one or more fjd-worker threads, like this:
$ fjd-recruiter hire [<number of workers>]
Per default, this starts n-1 worker threads, where n is the number of CPUs on your machine.
Put jobs in the job queue. You do this by putting a configuration file per job in a designated directory (e.g. ~/.fjd/default/jobqueue, where ‘default’ could be changed to a specific project name). Here is an example job:
[control] executable: python example/ajob.py logfile: logfiles/job0.dat [params] param1: value0
I’ll talk about the details of these job files below and there is a full example as well.
Then, start a dispatcher:
Now the fjd-dispatcher assigns jobs to fjd-worker threads who are currently not busy.
You can configure a number of hosts in your network and how many workers should be running on each (see an example of this below).
First, you need to have python 2.7, which the default python on almost all systems these days (note: python 3.x support is not there yet, but close; see issue 10 on github). Then:
$ pip install fjd
If you do not have enough privileges (look for something like “Permission denied” in the output), install locally (for your user account only):
$ pip install fjd --user
If you do not have pip installed (I can’t wait for everyone running Python 3.4), I made a small script, which should help to install all needed things. Download it and make it executable:
$ wget https://raw.github.com/nhoening/fjd/master/fjd/scripts/INSTALL $ chmod +x INSTALL
Now you can install system-wide:
or, if you do not have root privileges, you can also install locally:
$ source INSTALL --user
Note - If you installed locally, this should be added to your ~/.bashrc or ~/.profile file:
Note - Installing locally could be the better choice, actually, because it might save you from installing fjd on each machine you want to use. If they all share the home directory, they will all know about fjd once you are logged in.
How does fjd work, in a nutshell?
Small files in your home directory are used to indicate which jobs have to be done (these are created by you) and which workers are available (these are created automatically). Files are also used by fjd to assign workers to jobs.
This simple file-based approach makes fjd very easy to use.
For CPUs from several machines to work on your job queue, we make one necessary assumption: We assume that there is a shared home directory for logged-in users, which all machines can access. This setting is very common now in universities and companies.
A little bit more detail about the fjd internals: The fjd-recruiter creates worker threads on one or more machines. The fjd-worker processes announce themselves in the workerqueue directory. The fjd-dispatcher finds your jobs in the jobqueue directory and pairs a job with an available worker. It then removes those entries from the jobqueue and workerqueue directories and creates a new entry in jobpods, where workers will pick up their assignments.
Then, the dispatcher calls your executable script and passes the file that describes the job to it as parameter on the shell. Your script simply has to read the job file and act accordingly.
All of these directories mentioned above exist in ~/.fjd and will of course be created if they do not yet exist.
A job file should adhere to the general INI-file standard. fjd only has some requirements for the control section, in which you specify which command to execute and where results should go. Here is an example:
[control] executable: python example/ajob.py logfile: logfiles/job0.dat [params] param1: value0
Your executable (the “job”) gets this configuration file passed as a command line argument, so this would be called on the shell:
python example/ajob.py <absolute path to the job file>
This way, it can see for itself in which logfile to write to. In addition, you can put other job-specific configuration in there for the executable to see, as I did here in the [params]-section (I repeat: only the [control]-section is required by fjd).
Take care to get the relative paths correct (or simply make them absolute): If the paths are relative, they should be relative to the directory in which you start the fjd-dispatcher.
To add this job to the job queue, we would place that file into ~/.fjd/default/jobqueue and the fjd-dispatcher will find it there.
Note You can specify a project name (example below) and then “default” would be replaced by that.
An example (on your local machine)
You can see how it all comes together by looking at the simple example in the example directory on github. There is one script that represents a job (example/ajob.py) and one that creates ten jobs similar to the one we saw above and puts them in the queue (example/create_jobs.py).
To run this example, create jobs using the second script, recruit some workers and start a dispatcher. Then, lean back and observe. We have a script that does all of this in run-example.sh:
#/bin/bash python create_jobs.py fjd-recruiter hire 4 fjd-dispatcher
And this is output similar to what you should see:
$ cd fjd/example $ ./run-example.sh [fjd-recruiter] Hired 4 workers in project "default". [fjd-dispatcher] Started on project "default" [fjd-dispatcher] Found 10 job(s) and 4 worker(s)... [fjd-dispatcher] Found 6 job(s) and 1 worker(s)... [fjd-dispatcher] Found 5 job(s) and 2 worker(s)... [fjd-dispatcher] Found 3 job(s) and 1 worker(s)... [fjd-dispatcher] Found 2 job(s) and 3 worker(s)... [fjd-dispatcher] No (more) jobs.
You can cancel the fjd-dispatcher process now (i.e. hit CTRL-C).
And you’ll see the results, the log files written by our example jobs:
$ ls logfiles/ job0.dat job2.dat job4.dat job6.dat job8.dat job1.dat job3.dat job5.dat job7.dat job9.dat
Workers are Unix screen sessions, you can see them by typing:
$ screen -ls
and inspect them if you want. As attaching to screen sessions is cumbersome and fjd can also close them before you have a chance to see what went wrong (this is an option you can set, see next example below), fjd logs screen output to ~/.fjd/<project>/screenlogs (each screen has its own log file).
Here is an example log from a screen session of a worker:
$ fjd-worker --project default [fjd-worker] Started with ID nics-macbook.fritz.box_1382522062.31. [fjd-worker] Worker nics-macbook.fritz.box_1382522062.31: I found a job. [fjd-worker] Worker nics-macbook.fritz.box_1382522062.31: Finished my job. [fjd-worker] Worker nics-macbook.fritz.box_1382522062.31: I found a job. [fjd-worker] Worker nics-macbook.fritz.box_1382522062.31: Finished my job.
By the way, if screen sessions are running and you want them to stop, then you can always fire workers by hand:
$ fjd-recruiter fire
$ fjd-recruiter --project <my-project> fire
If you start a new dispatcher, it will first clean up (“fire”) old screen sessions.
Another example (using several machines in your network and a custom project name)
We can tell fjd about other machines in the network and how many workers we’d like to employ on them. To do that, we place a file called remote.conf in the project’s directory. Here is my file example/remote.conf: If you run this example, you’ll have to fill in names of machines in your particular network, of course:
[host1] name: localhost workers: 3 [host2] name: hyuga.sen.cwi.nl workers: 5
Normally, that directory is ~/.fjd/default. In this example, we tell fjd to use a different project identifier (this way, you could have several projects running without them getting into each other’s way, i.e. stopping one project wouldn’t stop the workers of the other and you wouldn’t override the first project if you start another). Here is the content of run-remote-example.sh, using the project identifier remote-example:
#/bin/bash python create_jobs.py remote-example cp remote.conf ~/.fjd/remote-example/remote.conf fjd-recruiter --project remote-example hire fjd-dispatcher --project remote-example --end_on_empty_queue
If you run this example, the output you’ll see should be similar to this:
$ cd fjd/example $ ./run-remote-example.sh [fjd-recruiter] Hired 3 workers in project "remote-example". [fjd-recruiter] Host hyuga.sen.cwi.nl: [fjd-recruiter] Hired 5 workers in project "remote-example". [fjd-dispatcher] Started on project "remote-example" [fjd-dispatcher] Found 10 job(s) and 8 worker(s)... [fjd-dispatcher] Found 2 job(s) and 4 worker(s)... [fjd-dispatcher] No (more) jobs. [fjd-recruiter] Fired 3 workers in project "remote-example". [fjd-recruiter] Host hyuga.sen.cwi.nl: [fjd-recruiter] Fired 5 workers in project "remote-example".
Note Unlike in previous example, this time I told the fjd-dispatcher process to fire workers (kill screen sessions) and terminate itself once it finds the queue of jobs being empty.
Note - If you normally have to type in a password to login to a remote machine via SSH, you’ll have to do this here, as well. You can configure passwordless logon by putting a public key in ~/.ssh/authorized_keys. For the shared-home directory setting we use fjd for, this makes a lot of sense, as you stay within your LAN anyway. In general, some SSH configuration can go a long way to ease your life, e.g. by connection sharing through the ControlAuto option. Search the web or ask your local IT guy.