Modernizing research computing with the new Longwood cluster

All users of O2 will be affected by this transition. This includes: 

• HMS faculty, staff, and students who run computational workloads on O2 
• Researchers at affiliate hospitals and the Wyss Institute that are using O2 
• Users of O2-connected services, including the O2 Portal, O2-hosted websites, and SQL databases 

The transition from O2 to the new Longwood cluster follows a phased approach: 

1. Beta testing period – Early adopters will get access to advanced GPU resources and test the new cluster and help identify issues before broader access. 
2. Preparation period – All O2 users get access to the Longwood cluster to test their workflows and confirm that results match. Testing your workflows during this period helps avoid delays after your migration date. 
3. Migrations –  After HMS IT migrates your lab’s group folders, you continue your work on Longwood. You receive advance notice before your migration date. 
4. O2 retirement – After all migrations are complete, O2 will be decommissioned. 

The new Longwood cluster uses software and job-scheduling methods similar to those of O2. However, there are differences in how you access resources and manage workflows. Documentation and training will be available to support your transition. 

Now: 

• Familiarize yourself with your current workflows and dependencies. 
• Identify data in home folders or scratch storage that you want to retain (these may not be automatically migrated). 
• Watch for announcements about beta testing opportunities 

During beta testing: 

• Report any issues to rchelp@hms.harvard.edu so we can address them before broader access 

During the preparation period (all O2 users): 

• Test your workflows on the new Longwood cluster to confirm that results match what you get on O2 
• Resolve any differences before your migration date to avoid delays 
• Contact rchelp@hms.harvard.edu if you encounter issues 

When the migration period is announced: 

• Plan for a brief pause in work during your transition window. 
• Ensure any critical data is backed up before your migration date. 
• Follow the detailed migration instructions and checklists that we will provide later, when the migration schedule is finalized. 

This timeline will be updated with specific dates as they are confirmed.

Several services are connected to O2 and will transition to the new Longwood cluster:

• O2 Portal – The graphical interface for submitting jobs. A Longwood portal will be available when the new cluster launches.
• O2 computational web sites – Computational sites hosted through O2. We will provide migration details separately.
• SQL databases – Databases connected to O2. We will provide migration details separately.

If you use any of these services, watch for additional communications specific to your needs.

Support resources for this transition will include:

• Documentation on migrating workflows from O2 to Longwood
• Training sessions and workshops
• Direct assistance from Research Computing Consultants (RCCs)

For questions about the Longwood cluster transition, contact the HMS IT Research Computing team at rchelp@hms.harvard.edu.