FASSE Cluster – FAS Secure Environment
- 1 OVERVIEW
- 2 HRDSP REQUIREMENTS
- 3 ACCOUNT SIGNUP
- 4 NEW ACCOUNTS
- 5 SLURM and PARTITIONS
- 6 VDI ACCESS
- 7 COMMAND LINE ACCESS
- 8 FASSE FAQ
The FAS Secure Environment (FASSE) is a secure multi-tenant cluster environment to provide Harvard researchers access to a secure enclave for analysis of sensitive datasets with DUA’s and IRB’s classified as Level 3. All servers in the FASSE environment are physically located inside an access-controlled data center. We have implemented security controls and access control lists to restrict access. Access to the cluster is restricted via a Virtual Private Network (VPN) and only authorized users/groups will be added to the FASSE VPN realm. We provide different storage tiers based on project needs. Please review storage options.
Note: As this is a secure environment, your home folders on FASSE is separate from any home folder you might have on the FASRC (Cannon) cluster. Data from the secure level 3 (FASSE) environment should not be transferred into level 2 space (Cannon).
In order to have a FASSE environment created for a project, the project owner or PI must first satisfy the HRDSP application requirements. FASRC is required by the university to review any documents (DUA/DAT/IRB) before a new FASSE project is created and/or any data is copied to the cluster. This information will also help FASRC determine how the environment should be set up, who the contacts are, and how project group names should be constructed.
FASRC cannot advise you on this, please contact VPR for assistance and guidance.
HRDSP Applications Summary and Order of Reviews: https://vpr.harvard.edu/files/ovpr-test/files/hrdsp_applications_summary_with_order_of_reviews.pdf?m=1595514758
If you have completed the above process, you can reach out to us at firstname.lastname@example.org with your project details. Please include the information shown in the following table (click + to expand):
Please provide the following in your email. Please do not submit a ticket before completing the HRDSP requirements noted above.
- Have you followed the HRDSP Applications Summary and Order of Reviews and completed the necessary Data Safety Application and/or EST/IRB Application process? (Yes/No):
- Who is listed as the primary project contact?:
- Please provide any project IDs you have/have received:
- Please provide a short name for the project (will be used in generating the group, storage, etc.) Alternately, we will extrapolate a short name from the HRDSP submissions:
- Who is the Harvard PI (must be faculty member or a designated non-faculty PI account) responsible for you or for the project?:
- Are additional team members listed on your IRB/DUA who should be added to the environment once it is set up? (Please list and whether they have FASRC accounts already):
- What are your platform requirements? FASSE Level 3 cluster, Cannon Level 2 cluster, Linux VM, Windows VM
- Do you require any of the following (Yes/No)? - One or more GPUs: - Large amounts of RAM (greater than ~ 120GB): - Data must be encrypted at rest (Tier 2 CEPH):
- How much storage do you estimate the project will need initially? (This is only an estimate to help FASRC in planning.) Unit is Terabytes (TB):
- FASRC offers several tiers of storage. [link] Provide your best guess as to the type of storage your project will need.: Tier 0/1/2
- If you are a Harvard faculty PI (or non-faculty researcher with explicit PI rights) you should have or request an account as a Faculty/Non-Faculty PI with FASRC. How Do I Get a FAS Research Computing Account?
- If you are not a Harvard faculty PI (or non-faculty researcher with explicit PI rights), you will need to have/sign up for an account under a Harvard PI/sponsor. How Do I Get a FAS Research Computing Account?
Before you can access the FASSE cluster you need to request a Research Computing account, selecting your PI as your sponsor (in this case, this is a Harvard faculty PI [or in some cases, a non-faculty researcher with PI rights], not necessarily the person listed on an IRB or DUA). See How Do I Get a Research Computing Account for instructions if you do not yet have an account. If your Harvard PI does not exist, please direct them to this same page and the directions in the previous paragraph.
If you have an existing account or have already completed the following three steps, you can skip this section. But please note the FASSE VPN realm (@fasse) noted below. You must connect to this realm to access any FASSE resources.
Once you have your FASRC account, you will receive an email with the same information as below, but step one is to set your password. This will be done using your email address and our password reset system.
To access FASSE and most FASRC services, including the FASRC VPN, you will need your personal FASRC OpenAuth two-factor (2FA) token. This can be set up on your smartphone using an app or downloaded as a Java applet to run on your desktop/laptop.
In order to access any secure system or environment in FASRC, you will need to connect to the FASRC VPN. The FASRC VPN is separate from other Harvard VPNs you may already be using. To connect to a FASSE environment, you will connect to the FASRC VPN (vpn.rc.fas.harvard.edu) using the @fasse realm (ex. – jharvard@fasse), your FASRC password, and your OpenAuth 2FA code.
SLURM and PARTITIONS
To manage the workload on the cluster we use SLURM. Following are the partitions available on the FASSE cluster.
To use slurm on the cluster, please check here
|Partition||Number of Nodes||Cores per Node||CPU Core Types||Mem per Node (GB)||Time Limit||Max Jobs||Max Cores||MPI Suitable?||GPU Capable?|
|fasse||24||48||Intel "Cascade Lake"||184||7 days||none||none||yes||No|
|fasse_gpu||4||32||Intel||124||7 days||none||none||yes||Yes (4 K80/node)|
|test||4||48||Intel "Cascade Lake"||184||8 hours||none||none||yes||No|
VDI, aka OpenOnDemand or OoD, is a virtual GUI interface that provides everything from pre-built apps to interactive command line access within a familiar desktop-like environment. The FASSE VDI is available when connected to our @fasse VPN realm, through your web browser. Please visit to access the service: https://fasseood.rc.fas.harvard.edu
See our FASSE VDI documentation for further information. [Link Pending]
COMMAND LINE ACCESS
Command-line access is also available for those who need/want to run jobs using a CLI. Login nodes for FASSE can be accessed by SSH at fasselogin.rc.fas.harvard.edu
See our FASSE CLI documentation for further information. [Link Pending]
Interim Documentation: See the very similar main cluster doc in the interim
Level 3 and other sensitive files and data stored within the secure environment should never be transferred to storage on the FASRC main cluster or to outside storage which is not designed and approved to house secure data.
FASSE secure storage shares should be accessible via Globus to allow you to transfer your data.
Local Scratch on FASSE Nodes
Jobs on FASSE nodes should use /tmp for private temporary space. That space is accessible only by that job and not other jobs on the same node. We do not recommend using /scratch as any data there will be visible to any other job running on the same node.
Global scratch is available at
/n/holyscratch01 or using the
FASSE global scratch has the same 90-day retention policy. For policy details and more on the scratch variable, see: Scratch Policy
When logging into FASSE you will have a home directory that resides only on FASSE. When logging into the main cluster, you will find a different home directory. So bear this in mind if you do switch between the two.
Accessing the internet while connected to the FASSE VPN realm (@fasse) and any FASSE nodes is managed by a proxy.
This should be a global environment variable, but some applications may require you to manually provide them.
The proxy settings are:
For HTTP: http://rcproxy.rc.fas.harvard.edu
For HTTPS: https://rcproxy.rc.fas.harvard.edu
Port: Both use port 3128
Please note: Our proxy does not allow all traffic, but should allow access to things necessary for your work. For web browsing, your browser should work if set to ‘Use system proxy settings’ (Auto-detect proxy’ may also work). If this does not work automatically, you will need to manually add the proxies to your browser. You will need to disable this when not on the VPN.
HTTP Proxy: http://rcproxy.rc.fas.harvard.edu
HTTPS Proxy: https://rcproxy.rc.fas.harvard.edu