Disk areas
CSC supercomputers have three main disk areas: home, projappl and scratch. In addition to these disk areas visible to all compute and login nodes, each node has a local temporary disk area that is visible to the particular compute node during a batch job or shell session, only. Please familiarize yourself with the areas and their specific purposes. The disk areas for different supercomputers are separate, i.e. home, projappl and scratch in Puhti cannot be directly accessed from Mahti. Also a more technical description of the Lustre filesystem used in these directories is available.
CSC does not backup your data!
None of the disk areas are automatically backed up by CSC! This means that data accidentally deleted by the user cannot be recovered in any way. To avoid unintended data loss, make sure to perform regular backups to, for example, Allas. See also the allas-backup tool.
| Owner | Environment variable | Path | Cleaning | Automatic backup | |
|---|---|---|---|---|---|
| home | Personal | ${HOME} |
/users/<user-name> |
No | No |
| projappl | Project | Not available | /projappl/<project> |
No | No |
| scratch | Project | Not available | /scratch/<project> |
180 days on Puhti | No |
These disk areas have quotas for both the amount of data and total number of files:
| Capacity | Number of files | |
|---|---|---|
| home | 10 GiB | 100 000 files |
| projappl | 50 GiB | 100 000 files |
| scratch | 1 TiB | 1 000 000 files |
LUE
To easily check the amount of data and number of files within a given folder on
the parallel file system, please consider using the LUE
tool. This tool is significantly faster than tools like stat or du and causes
much less load on the file system.
Quotas and cleaning
While it is possible to apply for increased quotas, we recommend that you always first ensure that the data you have stored on the shared file system is really needed and in active use. Unused data should be moved to e.g. Allas. A general tutorial on managing and cleaning data on Puhti and Mahti disks is also available.
Home directory
Each user has a home directory ($HOME) that can contain up to 10 GB of data.
The home directory is the default directory where you begin after logging in to CSC supercomputers.
However, typically you should change to your project's scratch directory when working because
the home directory is not intended for data analysis or computing. Its purpose is to store
configuration files and other minor personal data. A home directory exceeding its capacity
causes various account problems.
The home directory is the only user-specific directory in supercomputers. All other directories
are project-specific. If you are a member of several projects, you also have access to several
scratch or projappl directories, but still have only one home directory.
Scratch directory
Each project has by default 1 TB of scratch disk space in the directory /scratch/<project>.
This fast parallel scratch space is intended as temporary storage space for the data that is used in the supercomputer. The scratch directory is not intended for long-term data storage. To ensure that the disks do not fill up CSC will regularly delete files that have not been accessed in a long time. In Puhti the current policy is to remove files that have not been accessed for more than 6 months. In Mahti a similar cleaning procedure will be introduced, but is not yet active. See Usage policy page for details on the current policy.
Make sure to consult our tutorial for tips and guidelines on how to
manage your data on scratch.
Projappl directory
Each project has also a 50 GB project application disk space in the directory
/projappl/<project>.
It is intended for storing applications you have compiled yourself, libraries etc. that you are sharing within the project. It is not a personal storage space but it is shared with all members of the project team. Note that no files in this folder will be removed automatically.
It is not intended for running applications, so please run them in scratch instead.
Using scratch and projappl directories
An overview of your directories in the supercomputer you are currently logged on can be displayed with:
The above command displays all scratch and projappl directories you have access to.
For example, if you are a member in two projects, with unix groups project_2012345
and project_3587167, then you have access to two scratch and projappl directories:
[kkayttaj@puhti ~]$ csc-workspaces
Disk area Capacity(used/max) Files(used/max) Project description
----------------------------------------------------------------------------------
Personal home folder
----------------------------------------------------------------------------------
/users/kkayttaj 2.05G/10G 23.24k/100k
Project applications
----------------------------------------------------------------------------------
/projappl/project_2012345 3.056G/50G 23.99k/100k Ortotopology modeling
/projappl/project_3587167 10.34G/50G 2.45/100k Metaphysics methods
Project scratch
----------------------------------------------------------------------------------
/scratch/project_2012345 56G/1T 150.53k/1000k Ortotopology modeling
/scratch/project_3587167 324G/1T 5.53k/1000k Metaphysics methods
Moving to the scratch directory of project_2012345:
Please note that not all CSC projects have Puhti/Mahti access, so you may not
necessarily find a scratch or projappl directory for all your CSC projects.
Note
The scratch and projappl directories are shared by all the members of the
project. All new files and directories are also fully accessible for other
group members (including read, write and execution permissions).
If you want to restrict access from your group members, you can reset the permissions
with the chmod command. Setting read-only permissions for your group members for
the directory my_directory:
As mentioned earlier, the scratch directory is only intended for processing data.
Any data that should be preserved for a longer time should be copied to the Allas
object storage server. Instructions for backing up files from CSC supercomputers to
Allas can be found in the Allas guide.
Moving data between supercomputers
Data can be moved between supercomputers via Allas by first uploading the data in one supercomputer and then downloading in another supercomputer. This is the recommended approach if the data should also be preserved for a longer time.
Data can also be moved directly between the supercomputers with the rsync command.
For example, in order to copy my_results (which can be either file or directory)
from Puhti to the directory /scratch/project_2002291 in Mahti, one can issue in
Puhti the command:
See Using rsync for more detailed instructions for rsync.
Increasing quotas
You can use the MyCSC portal to manage quotas of the scratch and projappl
directories.
Remember that even after the quota is increased, the planned automatic cleaning process
will continue removing idle files from the scratch directory. Data that is not under
active computing should be stored in the Allas storage service.
Remember also that you can increase these values only to some extent. Especially regarding
the number of files, you should reconsider your data workflow if it requires that tens
of millions of files are stored in the scratch area.
Info
To find out how much data/files you have on the disk, please use our LUE
tool which is much more performant than standard
tools such as stat or du.
Temporary local disk areas
If the application depends on the use of temporary files, the suitability of
the filesystem may have a large effect on the performance of the application,
see section Mind your I/O - it can make a big difference in the Performance
checklist.
Please note that some applications use temporary files "behind the scenes". Usually these
applications read some environment variable that points to a suitable disk area, such as
$TMPDIR.
Some nodes have local disks that can be used to speed up your work when the temporary files are only needed within a single login- or compute node.
Login nodes
Each of the login nodes have 2900 GiB of fast local storage. The storage is located under
$TMPDIR and is separate for each login node.
The local storage is good for compiling applications and performing pre- and post-processing that require heavy I/O operations, for example packing and unpacking archive files.
Note
The local storage is meant for temporary storage and is cleaned frequently. Remember to move your data to a shared disk area after completing your task.
Compute nodes with local SSD (NVMe) disks
Jobs running in the I/O- and GPU-nodes in Puhti and GPU-nodes in Mahti have local fast storage
available. In interactive batch jobs launched with sinteractive,
this local disk area is defined with environment variable $TMPDIR and in normal batch jobs
with $LOCAL_SCRATCH. The size of this storage space is defined in the batch job resource request.
Different nodes have different amounts of disks, see Puhti technical details
for a detailed list of all node types. In normal compute nodes, there are 1490 GiB and 3600 GiB
disks. In big memory nodes there are 1490 GiB and 5960 GiB disks, and in GPU-nodes there are
3600 GiB disks. To save resources, and to ensure your jobs do not queue for resources for too
long, it is a good idea to only reserve what you actually need.
These local disk areas are designed to support I/O intensive computing tasks and cases where you
need to process large amounts (over 100 000) of small files. These directories are cleaned once
the batch job finishes. Thus, in the end of a batch job you must copy all the data that you want
to preserve from these temporary disk areas to scratch directory or to Allas.
For more information see creating job scripts.
Compute nodes without local SSD (NVMe) disks
In Puhti we simply recommend using compute nodes with NVMe disks ($LOCAL_SCRATCH) for the
applications that require temporary local storage.
In Mahti, with most compute nodes without local NVMe disks, it is possible to store a relatively
small amount of temporary files in memory. In practice, the applications can use the directory
/dev/shm for this, for example by setting export TMPDIR=/dev/shm. Please note that the use
of /dev/shm consumes memory, so less is left available for the applications. This may lead to
applications running out of memory sooner than expected and failing in the compute node, but
this usually does no other harm. The plus side is that if it works, it should be fast. In Puhti
however, where applications from multiple users can share the same node, running out of memory
by filling up /dev/shm will crash other users applications, too. It is thus not recommended to
use /dev/shm in Puhti at all.