Prerequisite for using the JUROPA / HPC-FF system is to successfully pass through the application process. All details can be found at the dedicated website for the JUROPA / HPC-FF cluster (dispatch, english version). Specific information for HPC-FF users is available at HPC-FF Support.
Please note that the operation of the HPC-FF partition ends on June 30, 2013. After this date, only access to existing HPC-FF data will be possible. Login to nodes hpcff or juropagpfs for this purpose.
The Juropa/HPC-FF system is split into two partitions: JUROPA and HPC-FF. These two partitions include Compute nodes as well as Login nodes. Hence, depending on your membership to one of these groups, a different set of Login nodes has to be used:
|Group||Generic Name||Specific Node Names||Domain||Login Method|
|HPC-FF||hpcff||hpcff01, hpcff02, hpcff03||fz-juelich.de||ssh|
|JSC||juropa||juropa01, juropa02, juropa03, juropa04, juropa05, juropa06, juropa07||fz-juelich.de||ssh|
Users outside the FZ campus have to use fully qualified names including the domain name. When using the generic name, a connection is established to one of the Login nodes in the node pool. Initiating two logins in sequence may lead to sessions on different nodes. In order to force the session to be started on the same nodes use the specific node names instead.
Users can't login by suppling username/password credentials. Instead, password free login based on SSH key exchange is required.
The public/private ssh key pair has to be generated on the workstation you are using for accessing JUROPA / HPC-FF. On Linux or UNIX-based systems, the key pair can be generated by executing
ssh-keygen -t [dsa|rsa]
Please protect the ssh key with a non-trivial pass phrase to fulfill the FZJ security policy. (For further details on ssh key generation see FAQ: How to generate ssh keys)
The generated public ssh key contained in the file id_dsa.pub or id_rsa.pub on your workstation must be uploaded through the WEB interface when initially applying for a userid on JUROPA/HPC-FF. In case your ssh key got lost or needs to be updated, you can use the dispatch ssh-key upload page, to upload a new key. In either case the key will be put to the file
on the JUROPA/HPC-FF system.
Note: Please make sure that your home directory on JUROPA/HPC-FF is not open for write access for other users or user groups, otherwise .ssh will not work.
After the ssh key has been generated and uploaded, a typical login looks like
(see examples above). Here, juropa is the generic name which will establish a connection to one of the Login nodes from the set juropa01, ..., juropa07. Accordingly, for the generic login to hpcff, one of the nodes of hpcff01, ..., hpcff03 will host the session.
Please note: Too many accesses (ssh or scp) within a short amount of time will be interpreted as an intrusion and will lead to automatically disabling the originating system at the FZJ firewall. For transferring multiple files in a single scp session, the -r option can be used, which allows to transfer a whole directory.
If X11-based graphical tools are to be used on JUROPA/HPC-FF, it may be necessary to enable X11 forwarding in the file /etc/ssh/config or $HOME/.ssh/config on your workstation:
| PubkeyAuthentication yes|
or to use the -X ssh option (possibly combined with -A for authentication forwarding)
For all HPC-FF and JSC users of JUROPA/HPC-FF, an additional set of front-end nodes is available, which give access to the GPFS filesystem. The purpose of these nodes is to provide the most convenient way to copy files between the GPFS filesystems of JUGENE and the Lustre filesystems of JUROPA/HPC-FF. Furthermore, they provide an interface to the GPFS-based tape archiving system of FZJ. The following table provides the login details:
|Generic Name||Specific Node Names||Domain||Login Method|
|juropagpfs||juropagpfs01, juropagpfs02, juropagpfs03, juropagpfs04, juropagpfs05||fz-juelich.de||ssh|
Note: The GPFS filesystems are only available on the GPFS nodes, not on the compute nodes of JUROPA/HPC-FF. This means, they can't be accessed by batch jobs.
Although the GPFS nodes provide essentially the same login and programming environment as the login nodes, users might experience performance degrations due to data transfer happening on the GPFS nodes. Therefore, the recommendation is, to use the GPFS nodes only, if the specific features of the GPFS nodes are needed. (For more details on specific properties of the GPFS nodes see Memory Optimisation.)
The Login Shell on each system points to system's favored shell:
JUROPA -> bash
Users are not allowed to change the Login Shell, but they can switch to a personal shell within the login process. User's will find a template within the initial FZJ $HOME/.bashrc.
Home directory ($HOME)
Home directories reside in the Lustre filesystem. In order to hide the details of the home filesystem layout the full path to the home directory of each user is stored in the shell environment variable $HOME. References to files in the home directory should always be made through the $HOME environment variable. The initialization of $HOME will be performed during the login process.
Other important storage locations are the scratch filesystem ($WORK), GPFS home directories ($GPFS_HOME) and the archive filesystems ($ARCH). The following table summarizes the information:
|$HOME||Lustre filesystem||User's home directory|
|$WORK||Lustre filesystem||Scratch filesystem for temporary data|
|$GPFS_HOME||GPFS filesystem||User's home directory on the JUGENE cluster|
|$ARCH||GPFS filesystem||User's storage location for archiving on tape|
It is highly recommended to access files always with help of these variables. The values of these variables are automatically set during login.
Programming environment, modules and libraries
The persistent settings of the shell environment are governed by the content of .bashrc , .profile or scripts sourced from within these files. Please use these files for storing your personal settings.
The common programming environment is maintained with the concept of software modules in directory /usr/local. The framework provides a set of installed libraries and applications (including multiple-version support) and an easy to use interface (module command) to set up the correct shell environment (see also: module setup and compiler).