Quick Start IPSL Mesocenter using guide
Accessing the clusters
Connection to IPSL Mesocenter is made by SSH via the clusters head nodes.
Clusters head nodes are WorldWide Open via ssh but only with RSA key ( no password )Head nodes adresses by cluster :
- IPSL-SU, ciclad:
- ciclad.ipsl.upmc.fr
- ciclad2.ipsl.upmc.fr
- IPSL-X, climserv:
- camelot.ipsl.polytechnique.fr
- loholt1.ipsl.polytechnique.fr
- loholt2.ipsl.polytechnique.fr
- IPSL-G, hal: (GPU small Cluster)
- hal0.obs.uvsq.fr
To connect to a specific cluster, head node, you simply execute the command
ssh your_mesocenter_login@head_node_name
Your mesocenter login grants your access to all IPSL clusters.
see more on Head nodes page
Working and Datasets spaces
User and Projects spaces
The working spaces are divided in:
- User spaces
$HOME
: user home folder- Data space : user data folder
- Project spaces : which depends on the projects your involved in.
see more on About Working spaces page
Dataset space
Common data sets are accessible via
/bdd
.
see more on About Data sets page
Module command
IPSL mesocenter ( ciclad and climserv ) uses
module
command for software initialization
see more on: Module command page
Python environment
There is many python distribution on IPSL Mesocenter we recommand to use anaconda distribution for python:
- 2.7 use the module
python/2.7-anaconda50
- 3.6 use the module
python/3.6-anaconda50
[xxxxxx@ciclad-ng ~]$ module avail python
---------------------- /etc/modulefiles/Products ----------------------
python/2.7.3-epd7 python/2.7-anaconda50
python/2.7.6-canopy-1.3(default) python/3.4-anaconda3
python/2.7-anaconda python/3.6-anaconda50
python/2.7-anaconda44
see more on: Python environment page
Job Manager
The job manager allows you to submit your analisys to be executed on the computes nodes of the cluster. Each submission is refer as a 'job'.
see more on: About Job Manager page