Skip to content

Latest commit

 

History

History
47 lines (33 loc) · 3.48 KB

gricad.md

File metadata and controls

47 lines (33 loc) · 3.48 KB
Raw

How to access data from GRICAD

  • If you don't have an account, create it here (select chercheur in contract type if you have a permanent position or are a post-doc, doctorant if you are PhD) and ask to join the pr-sasip project.
  • Read carefully the computation center documentation
  • Once logged to dahu (CPU) or bigfoot (GPU), you have access to the storage at : /summer/sasip

Configuration of your access to gricad

You can set up a direct access to dahu via this procedure so that ssh and scp takes only one line of command, a simplified version is reported below :

  • Check that you should be able to connect to the bastions : ssh <your-login>@rotule.u-ga.fr and ssh <your-login>@trinity.u-ga.fr (replace with your perseus account)
  • Once connected on one of the bastions, check that you can connect to dahu (for CPU) : ssh dahu or bigfoot (for GPU) ssh bigfoot
  • To connect directly to dahu or bigfoot, add the following lines to your .ssh/config :
Host *
  ServerAliveInterval 30
  AddKeysToAgent yes
  IdentityFile ~/.ssh/id_rsa
  ForwardX11 yes
  TCPKeepAlive yes
 
Host *.ciment
  User <your-login>
  ProxyCommand ssh -q <your-login>@access-gricad.univ-grenoble-alpes.fr "nc -w 60 `basename %h .ciment` %p"
  • Now you should be able to connect to dahu directly with ssh dahu.ciment or to bigfoot with ssh bigfoot.ciment
  • To connect without typing your password, you can set up a SSH key :
    • on your local machine create the ssh key : ssh-keygen
    • copy the public key to the bastions, it is the file in .ssh that ends with .pub, it can be id_rsa.pub or id_ed25519.pub :

ssh-copy-id -i .ssh/KEY.pub <your-login>@rotule.u-ga.fr: and ssh-copy-id -i .ssh/KEY.pub <your-login>@trinity.u-ga.fr: and to the clusters : ssh-copy-id -i .ssh/KEY.pub <your-login>@dahu.ciment: or ssh-copy-id -i .ssh/KEY.pub <your-login>@bigfoot.ciment:

How to share a dataset on the SUMMER storage

Once you connected to dahu, go to /summer/sasip

  • put your data in a repository that has a meaningful name (ex ERAinterim-2012-2020-3h) in the right subdirectory (model-configuration, model-forcing, model-outputs, observations)
  • if needed, change the owner and rights to your directory so that everyone in the pr-sasip group can access it (like drwxrwxr-- 4 alberta pr-sasipfor instance)
  • provide a readme file that describes what, who, where, when the data was produced inside the repository (like this one for instance)
  • if you are the producer of the dataset, try as much as possible to follow the FAIR principles that SASIP committed to endorse (rich metadata, doi, etc ...)
  • report the same informations on the catalog on github : create a md file in the right subdirectory of github (ex forcings/your-data.md) and make a link to it in forcings.md in this example
  • advertise that your data is available among SASIP participants using participants(at)sasip-climate.com mailing list !