How to monitor running / pending jobs
The procedure below will help you find your running job, locate the results file, and view the results–even if the job is still running!
1. To see a list of running / pending jobs
You’ll need to login at the command line to the cluster.
To see a list of running/pending jobs you can run the command: “qstat -f” – if you are using SGE 6.1 or later there is a behavior change where “qstat -f” will now only show you your own jobs (in all previous revs of SGE, qstat -f would show jobs belonging to all users). For SGE 6.1 systems you need to use “qstat -f -u ‘*’” to explicitly list jobs from all users on the system. I can show you how to make this behavior the default if you want to bake it into the system before your expected users start using the system in earnest.
When you run “qstat -f” or “qstat -f -u ‘*’ " you will see a text based status display of grid status
What you need to do is identify your running blast job in the qstat display. What you really want to learn is the SGE job ID assigned to the job.
2. To find the filename of your results file
Once you learn the job ID of your blast job, you can run the command “qstat -j <job ID>” and that will dump lots of information about your job.
Included in the “qstat -j” output will be data that lists the directory that the job is running out of. Once you learn that you can simply enter that directory and look at the output files.
3. To view your results file in real time
The “tail -f ./filename” command is a great way of watching a job in real time. It is convenient to use this procedure with BLAST jobs to see how they are progressing.
