Job details

When you run a job, a number of details need to be known. Such as the machine were you wish to run, and more. The job details section collects all this information and shows it to the user. If you select a particular queue from the Queue menu, the job details will be reset to the proper values as defined for that queue. However, you can override them by changing them afterwards for a particular job. You can also change the name of queue shown with the job information, click on the name to do this. The information is stored in the .pid file from the job.

Remote host

Name of the machine on which you wish to run your job. You should be able to connect to that machine using ssh, and the name as specified. If you wish to run on your local machine, leave this field empty or specify localhost.

Remote user

The username that you need to specify in the ssh command, if any. Typically, this is your username on the remote machine.

Remote job directory

On the remote machine, ADFjobs needs to set up your input files and run script, and needs to collect the results. For that purpose ADFjobs will make its own directories within the directory you specify. A typical value would be something like $HOME/jobs.

Run command

The command on the remote machine to be executed. In this command, $job will be replaced by the full path to your job script on the remote machine. Also, $options will be replaced by the contents of the Run options field. Typically this will be used to specify the number of CPUs, a time limit, or a batch queue name.

To run interactively, just enter "$job". To submit your job to a queue, specify the submit command (for example, qsub). Several examples may be found in the preconfigured queues.

Kill command

The command to use to kill a queued or running job. In this command $pid will be replaced by the process id or the job id, as determined when starting the job. For interactive jobs, killall $pid should work fine.

Job status command

This command will be used to determine if a job is still queued or running. If a job is no longer queuer or running, it should return an empty string. Anything else will server as indication that the job is alive. For interactive jobs ps -p $pid | grep $pid works fine.

System status command

The command to use to determine the system status. This might be uptime, or some qstat command for batch systems.

Prolog command

The command to execute at the beginning of the job script. Typically, this will be used to set up the environment properly. For example, you would source a script file to set all environment variables for ADF like ADFHOME, ADFBIN etc. This is especially important if you are working with different versions of ADF at the same time.

Epilog command

This is the command to run at the end of the job script. You can use it to copy save result files, or to perform some cleanup action.

Logfile extension

The extention for the logfile, should normally be logfile. If you use ADFjobs for other programs than ADF, you could specify a different value to monitor the progress of your other program.