Connection to login machine
To log on, you must :
- have a validated IP address on the login machine (see details in Login Account form)
- have an opened account on the CICADA cluster (see details in Login Account form)
- have a SSH client on your personal computer :
- For Microsoft Windows, you can use Putty : http://www.chiark.greenend.org.uk/~sgtatham/putty/download.html
- For Linux and Unix-like, you can use the native SSH client that comes with the Operating System
Type the following command :
If you have some disconnection issues, you can use the following command :
ssh email@example.com -o ServerAliveInterval=15
2 login machines are avalaible cicada1 and cicada2, you can connect on the one or the other and they are perfectly identical
user@client ~$ ssh firstname.lastname@example.org email@example.com's password: Last login: Fri Oct 12 08:57:53 2012 from client Welcome to the cluster cicada _ __ _____ (_)_____ ____ _ ____/ /____ _ / ___// // ___// __ `// __ // __ `/ / /__ / // /__ / /_/ // /_/ // /_/ / \___//_/ \___/ \__,_/ \__,_/ \__,_/ NICE university [user@cicada1 ~]$
At the first connection on one of these 2 login machines, you have to change necessarily your password via the command passwd.
You can use SSHFS to login to the machine : see SSHFS
So you can access your both personal directories :
- your home directory /home/user where you can store your programs and documents
- your personal directory for compute in /workspace/user where you can temporarily store your calculation datas
WARNING : up to now
- These 2 spaces have no disk quota, if users abuse, quotas will be implemented ASAP.
- For /home, the maximum available space is 25 TeraBytes and shared by all.
- For /workspace, the maximum available space is 260 TB and shared by all.
- NONE OF THESE 2 DIRECTORIES ARE SAVED : you must do your own backups on your client computer !
Some tips for a good use of your disk space :
- You can work in your /home/user when you develop and test your code.
- You must work in your /workspace/user when you launch computing jobs and want to store resulting datas.
- For an easier use of workspace, it is advisable to create a symbolic link in your home directory via the command :
cd ~ ln -s /workspace/user workspace
For informations, the bandwidth output of GPFS /workspace is about 4Go/s for read / write !