Connect to Carina
You can access Carina via ssh or the web interface.
Step 1 Connect to the Full-Tunnel Stanford VPN
This is the #1 reason for connection failure! For help with the VPN, consult the excellent UIT documentation.
Step 2 Connect to Carina
Choose a connection method, or use both
SSH Connection
ssh <sunetid>@c2-login.carina.stanford.edu
This will land you on a login node. You are now logged in, but the login node is not where you run your work.
Note: Login nodes are intentionally resource-limited. Each user session is capped at 512 MB of memory and 25% of a CPU core. These limits exist to keep the login node responsive for everyone. If you try to run a computation there, it will either run very slowly or be killed without a clear error message.
The login node is the right place to edit files, manage your data, and submit jobs to the cluster. For actual computation, you need Slurm.
Slurm? What is that? Slurm Docs.
Carina OnDemand
Carina OnDemand is based on the popular Open OnDemand platform, which is used across many Stanford research systems.
First visit error? If Carina OnDemand gives you an error about a missing home directory the first time you visit, please log in via ssh as described above. This will create the missing directory and allow Carina OnDemand to load.
Step 3 Find your files
Your $HOME is /home/users/<sunetid>.
Project folders are at /projects/<PI>/<projectID>/main.