- Using your terminal program, log into Odyssey
- (optional) If you have XQuartz (Mac) or Xming (PC), Log into Odyssey w/ X11 forwarding. Start a GUI clock sessions with the command
xclock &
- (optional) Log on to one of our NX machines (
rcnx01
orholynx01
) and start a terminal session
Transferring files with FileZilla
- Download the file at https://github.com/fasrc/DataC-HPC-genomics/blob/master/data/genome_data.zip?raw=true to your desktop
- Using FileZilla transfer this file to your Odyssey account home/login folder
- On your laptop's Desktop, unzip the genome_data.zip archive
- Using Filezilla, transfer the genome_data folder over to your home folder on Odyssey
(Mac/Linux) Transferring files with rsync
rsync
can transfer files within hosts or across hosts. The format is:
rsync (-options) source destination
rsync (-options) username@host:source destination
rsync (-options) source username@host:destination
where source and destination are any valid unix path
- In your terminal window from
Login and Access
, make a new directory 'transfer' in your home folder with the commandmkdir transfer
. - Using your terminal program, transfer the zip archive from your Desktop to the
transfer
folder on Odyssey - Using your terminal program, transfer the genome_data folder from your Desktop to the
transfer
folder on Odyssey
- See what directory you are in using
pwd
command. Compare this with your neighbor. Any differences? - You can display filesystems with the
df
command. Use this command with the following options: - df /
- df .
- df -h .
- What command would you use to see what other fileystems are available on Odyssey? ?
-
Issue the command
bowtie
. What happens? -
Issue the appropriate command to display what modules do you currently have loaded. What are they?
-
Using the appropriate command, find what versions of
bowtie
are on the cluster. -
Load the latest version of bowtie. What command did you use?
-
What modules do you have loaded now?
-
Issue the
bowtie
command now. What's different? -
Load in the legacy modules
-
Look for the versions of
bowtie
on the cluster again. What do you notice? -
(Bonus!) Use the text editor
nano
to modify your.bashrc
to opt-in for the new modules
- Issue the appropriate command to get an interactive session w/ 4 GB RAM for 60 minutes
- Start a MATLAB command-line session. What commands do you need to issue to do this?
-
(single core) Using
nano
or your favorite text editor, create a SLURM submission script to run thetest.py
example code. Use therun_test.sh
script as a starting point. -
Modify your script so that it includes the proper SBATCH directives, Lmod opt-in, module loads, and commands to run the test.py script.
-
Include a mail notification so that you know the script completed successfully.
-
Forward the SUCCESS email to your TF, and paste in the contents of the run_test.sh file. (DO NOT send this script file through the mail system, as it might be stripped out as a potential virus by mailserver virus scan software).
-
Bask in the joy that you've run your first cluster job!
-
Once your job completes successfully, how much RAM did your job use? What command did you use? Go back and modify your job script to use the appropriate amount of RAM.