A growing number of research projects are storing large amounts of data on NeSI systems. To better support this growth, as well as optimise the performance and availability of our filesystems, we are introducing new data management policies and best practices for our HPC facilities.
By adopting these measures to regularly audit, clean and manage the amount of data on our filesystems, we’ll ensure they remain high-performing and responsive to your research computing workloads and data science workflows.
Upcoming changes to data management processes for project directories
4-15 October 2021
The NeSI project filesystem is becoming critically full, however it is currently storing a large amount of dormant data that has not been accessed for more than 12 months. We need your help to free up space on the project filesystem as soon as possible. Please review the data you are currently storing in any
/nesi/project/ directories and delete or relocate any files that are no longer required for ongoing computational and/or analytics work on NeSI.
We have started regular audits of data stored in project folders, using the same format as our nobackup auto cleaning (described here). See the file
/nesi/project/<project_code>/.policy.test/scan485/latest.summary.txt for a summary of the number and size of files within each project that have not been accessed for more than 485 days (this is ~15 months, and is the draft auto cleaning timeframe under consideration for the project filesystem).
If you need assistance with this, contact Support and we’d be happy to help or answer questions.
18 October 2021
We will begin a limited roll-out of a new feature to automatically identify inactive files in
/nesi/project/ directories and schedule them for deletion. Generally, we will be looking to identify files that are inactive / untouched for more than 12 months.
A selection of active projects will be invited to participate in this first phase of the programme. If you would like to volunteer to be an early tester / participant, please contact Support. Otherwise, we will be in touch with projects directly to onboard them.
Insights from this initial phase will inform the criteria and processes of the programme prior to it being released to the broader user community.
Alongside this work, we will also adopt a new policy on how long inactive data may be stored on NeSI systems, particularly once a research project itself becomes inactive.
Starting in January 2022, we will expand the
/nesi/project/ directory data management programme to include all active projects on NeSI. Additional Support documentation and user information sessions will be hosted prior to wider implementation, to provide advance notice of the change and to answer any questions you may have around data lifecycle management.
Frequently asked questions
1) Why are you introducing these new data management processes?
We want to avoid our online filesystems reaching critically full levels, as that impacts their performance and availability for users. We also want to ensure our active storage filesystems aren't being used to store inactive data. This new data management feature for
/nesi/project/ directories will complement our existing programme of automatic cleaning of the /nobackup file system.
2) Can I check how much storage I’m currently using on NeSI systems?
You can query your actual usage and disk allocations at any time using the following command:
The values for 'nn_storage_quota' are updated approximately every hour and cached between updates.
3) Can I recover data that I accidentally delete from my /project directory?
Perhaps. We regularly make read-only copies of the file system and save them for up to seven days. For more information, refer to our File Recovery page.
4) Where should I store my data on NeSI systems?
How often will my team's HPC jobs be accessing the data?
How often will my team's HPC jobs be modifying the data?
Often (at least once every two months)
Store in your /nobackup/<projectcode> directory (but ensure key result data is copied to the persistent project directory).
Store in your /project/<projectcode> directory.
Apply for an allocation to use NeSI’s long-term storage service or store elsewhere (e.g. at your institution).
In general, the project directory should be used for reference data, tools, and job submission and management scripts. The nobackup directory should be used for holding large reference working datasets (e.g., an extraction of compressed input data) and as a destination for writing and modifying temporary data. The nobackup directory can also be used to build and edit code, provided that the code is under version control and changes are regularly checked into upstream revision control systems. The long-term storage service should be used for larger datasets that you only access occasionally and do not need to change in situ.
5) What should I do if I run out of storage space?
There are two tracked resources in the NeSI filesystem, disk space and inodes (number of files). If you run into problems with either of these, refer to this Support page for more information.
6) I have questions that aren’t covered here. Who can I talk to?
Contact Support. No question is too big or small and our intention is always to work with you to find the best way to manage your research data.
This page will continue to be updated in the coming months with more information. If you have questions at any time, don’t hesitate to contact Support.