Storage

Storage

Table of Contents:


Persistent Storage

userVolumes.png
This graphic shows the two types of persistent storage. Home Directories and Research Volumes.

Storage on the chip cluster environment is comprised of a personal home directory, shared research volume, and temporary/scratch storage. The first two types of storage are remotely hosted volumes from our Dell Isilon Storage System mounted with Network File System (NFS). With this, all of these volumes are able to be available across the chip cluster environment. Scratch storage is local to the allocated compute hardware you are using. It should be considered the fastest, but also most transient storage.


Home Directories

This is the home directory for your account, and will be your working directory immediately upon login. This directory should be sued for setting up your Unix user environment, as well as storing small amounts of data (eg., some small code, config files, etc). The storage space in /home/$USER is quite limited, only offering 500 MB of space. So utilize this space carefully.

Some common software packages will often attempt to write to the user home directory include conda, pip, and vscode, MATlab. These types of software all create “hidden” directories within the user home directory by default to store their data for operation. However, this can cause the home directory to be filled up quickly. To avoid this, please refer to the following links for managing home directory storage and avoiding those “hidden” directories from using all of your storage space.

How to check home directory usage on chip?

What to do if I run out of space in my chip home directory?

HPCF Research Volumes

On chip, the research storage is split into two volumes named after the research groups.

Location

Description

Location

Description

/umbc/rs/$GROUP/common

Shared storage for your research group

/umbc/rs/$GROUP/users/$USER

Personal storage in your research volume

This is where you will be storing majority of the data and projects being worked on.

Consider storing your personal data, model files, or experimental output in your /umbc/rs/$GROUP/users/$USER directory.

Consider storing group data or code in your /umbc/rs/$GROUP/common directory.

How to check research volume usage on chip?


Temporary Storage

On all machines, the /tmp directory can be used for storing small, temporary data. For example, this directory is commonly used when installing packages with pip or conda. For more information about the /tmp directory, please review the following webpage:

Why am I getting "OSError: [Errno 28] No space left on device" error ?

On all machines, the /scratch directory also exists. This storage is available on the compute nodes while jobs are running. This storage space is mainly intended to be temporary storage and provide extremely high I/O throughput to or from disk.

Because the /scratch space is local to the compute node, it is generally much faster than your /home/$USER directory or research volume (mounted via NFS). This enables it to provide much better latency when performing a large number of I/O operations. It is ideal for storing intermediate data files for running jobs, which can help reduce the load on the system-wide storage and local network.

For more information about the scratch space, refer to the following webpage:

Scratch Space/Temporary Storage Directory


Accessing Research Volumes

As stated previously, it’s in your best interest to avoid storing large datasets in your home directory. To avoid this conundrum, and encourage collaboration among the PI’s group, we give access to research volumes with large amounts of space.

By default, new PI groups start with 250GB, and all users in a PI group have access to these volumes. Each research volume comes with a “common” directory, which is where any user in the PI group can view/edit/run files. Additionally, when a new user is added, a private directory is made within the research volume that only the user has access to.

The path for this volume can be found at /umbc/rs/$GROUP, the common directory can be found under /umbc/rs/$GROUP}/common and the user directory can be found under /umbc/rs/$GROUP/users/$USER. We have also created aliases for you to quickly access these directories. You can view these aliases by typing the command alias into your command.

For example:

[regularUser1@chip ~]$ alias alias doit_ada='cd /umbc/ada/doit' alias doit_common='cd /umbc/rs/doit/common' alias doit_user='cd /umbc/rs/doit/users/regularUser1' alias sinfo='sinfo -M chip-cpu,chip-gpu -p 2024,2021,2018,general,gpu' alias squeue='squeue -M chip-cpu,chip-gpu'

Note: The user and common directories in the volume share the same space, so filling one up will also fill the other up. Please reference the diagram at the top of the page.


How Much Space Is Left?

You only have a limited amount of disk space, and in some cases, you will be sharing disk space with other users. Please be aware of what you are using.

You should be aware of your storage consumption in the following areas:

  • HPCF Research Storage: /umbc/rs/$GROUP/ and /umbc/rs/$GROUP/users/$USER/

  • Home Directory: /home/$USER/ - This directory only has 500 MB of storage.

Commands for checking usage:

  • du -h ./ - Lists the size of all the files in ./ directory

  • df -h ./ - Lists statistics about amount of free disk space on the specified location (./)

Refer to the following web page for more information on checking directory usage:

How to check home directory usage on chip?

How to check research volume usage on chip?

 

Back to Main Page