This is an execution platform for running fits on a number of arc files. It is not very general, but it works. Here is how

The system expects a structure like this

/
scripts/
2005/
2005-output/
2006/
2006-output/

Do NOT CREATE THIS STRUCTURE

first, run ./update-all.sh It will create the 2005-2011 folders, and populate them with the nessesary scripts. 

Then, in each of the 20xx folders, create a resource file. A resource file contains the URLs of the arc files to download and process.
It consists of lines like this
netarkro@sb-prod-bar-001:/netarkiv/0004/filedir/134163-152-20111103020156-00060-sb-prod-har-005.statsbiblioteket.dk.arc
netarkro@sb-prod-bar-001:/netarkiv/0004/filedir/134171-152-20111103034215-00035-kb-prod-har-001.kb.dk.arc
netarkro@sb-prod-bar-001:/netarkiv/0004/filedir/134118-152-20111102094559-00012-sb-prod-har-002.statsbiblioteket.dk.arc
netarkro@sb-prod-bar-001:/netarkiv/0015/filedir/134135-152-20111102124315-00002-kb-prod-har-004.kb.dk.arc

Then run the load.sh script in the 20xx folders, like this
./load.sh <resource-file> 4
This splits the file in 4 equal-sized chunks, each of which will be processed as a separate job. 
When you have done this for all the 20xx folders, go to the root dir again and run
./startAll.sh
This starts the specified number of jobs in each 20xx folder.
You can see the number of currently running processes by using 
./listRunning.sh

To see the progress, run
./status.sh
It shows the number of processed archives vs. the total number, and the numbers done in the near past.

The output for the 20xx run will be placed in the 20xx-output folder. This is tar.gz files containing the fits information for each resource in the arc file.

To stop all the jobs again, use 
./stopAll.sh
Because of the intricate system of processes spawning processes, it might be a few minutes before everything have died down. The stop command only kills the bash scripts, so spawned java commands will have to terminate normally. 

When running, the arc files are downloaded to
20xx/files and the fits output and extracted resources is temporarily stored in 20xx/temp. To clean this up, use
./cleanAll.sh
Do not do this while the jobs are running.