Data Science Facility Usage Policies¶
Establishes the responsibilities of Donald Danforth Plant Science Center (DDPSC) personnel or other authorized users for the protection and proper use of DDPSC Data Science Facility (Data Science) resources.
This policy applies to all users of Data Science resources and their supervisors. Any exceptions to this policy must be approved by your department head and Data Science.
The Data Science infrastructure is a shared resource and as such, the rules for acceptable use help to protect the integrity of the services and data for all users. Acceptable use means respecting other users’ rights, computing resources, and all license and contractual agreements.
- General policies
- Users must comply with all DDPSC policies and relevant software licensing while using computing resources.
- Data Science resources are for research purposes only. Commercial use must be set up as a separate activity and discussed with DDPSC and Data Science staff.
- Users may only access computers, user accounts, and files they are authorized to access.
- Research data on the Data Science infrastructure must be treated with care with regard to access, retention, sharing, publishing, and attribution. Principal Investigators or other responsible data custodians should be consulted for any questions related to such issues.
- Users are responsible for reporting unauthorized access to Data Science resources.
- Computing policies
- Users should not run programs on the login and HTCondor job scheduling server (stargate). All jobs should be submitted to the HTCondor cluster instead. Some small (e.g. text editing) or essential (e.g. ssh, screen, tmux, etc.) programs are generally acceptable. A monitoring service runs on these servers that will email users about processes that violate this policy. If you are unsure or feel you were emailed in error, please contact the Data Science staff.
- HTCondor interactive sessions will automatically log out after 2 hours of inactivity, but to free resources for others, please shut down these sessions as soon as possible when you are finished.
- It is the user’s responsibility to request appropriate resources (CPUs, memory, and/or disk) for HTCondor jobs. It is often not possible to be exactly accurate, but users should make their best effort or ask for help. Data Science staff will contact users who have jobs that are using excessive resources and reserves the right to end jobs that threaten the system or other users’ jobs.
- Users should not run large numbers of jobs simultaneously without testing first. A good rule of thumb is to first test running 1, then 10, then 100 jobs to make sure there are no unintended problems before launching a larger number of jobs.
- Users should monitor jobs running in the HTCondor cluster periodically to make sure they are running as expected and utilizing resources as requested.
- On the Jupyter notebook server, notebooks are capable of running/persisting in the background when the browser is closed or disconnect from the network. When finished working, please halt or shutdown notebooks to free resources for other users.
- Storage policies
- Storage policies apply to both the central file server and archival storage systems.
- Individual users and supervisors are responsible for data management. Minimally, data management practices must comply with the DDPSC policy on the Retention of and Access to Research Data.
- Research data and software stored in user and group directories is (or should be treated as) work product that is owned by the department and the DDPSC (or relevant entity). Personal data should not be stored on this system. Users must comply with the DDPSC non-disclosure agreement.
- The Data Science storage system is a shared filesystem, so copyrighted materials that do not allow sharing should not be stored on this system.
- File transfer via SFTP can be done using the login server (stargate). All other file transfer activity should be done on the dedicated file transfer server (chronos), or by submitting a job or interactive session request to the HTCondor cluster.
- Prohibited activities
- Violations of the copyright, trade secret, patent or other intellectual rights protections of any person or
company. This includes (but is not limited to):
- Installation or distribution of “pirated” software that is not appropriately licensed for use by DDPSC or the user.
- Unauthorized use of copyrighted material such as music, movies, unlicensed software and digitization and distribution of print media.
- Exporting software, technical information, encryption software/technology in violation of international or regional laws.
- Introduction of malicious programs into a DDPSC computer, network or server (e.g., virus, worm, trojan, ransomware, etc.).
- Revealing account information or passwords or allowing use of your account by others. If another user needs an account they should contact Data Science staff with permission from their supervisor.
- Using DDPSC resources to obtain, transmit, view or store material that is in violation of sexual harassment or hostile workplace policies or laws.
- Causing or aiding in security breaches or disruptions of network communication.
- Intentional destruction of intellectual property or work product. See 3.3 above.
- Using Data Science resources for personal projects or commercial work without prior approval.
- Violations of the copyright, trade secret, patent or other intellectual rights protections of any person or company. This includes (but is not limited to):
A user found to have violated any of the prohibited activities will be immediately reported to their supervisor and Human Resources for disciplinary action, up to and including termination of employment or relationship with DDPSC. If warranted by activities that cause significant or repeated disruption, a user’s account may be suspended until a meeting with Data Science staff can be arranged. For all other policy violations, Data Science staff will work with users for compliance and, if necessary, may involve the user’s supervisor.
- 2018-01-03: Initial text finalized.
- 2018-11-06: Updated login/HTCondor server name.
- 2019-09-12: Updated facility name.