Data center relocation

All services provided or facilitated by HMS IT will be disrupted from Friday, January 3, 2025, at 6:00 PM to Saturday, January 4, at noon EST (UTC-5).

HMS building access, basic internet access, and HUIT-provided services that leverage HarvardKey authentication will not be affected.

While we expect administrative systems to be available again by noon on January 4, disruptions to O2 computing services and attached storage will continue through Friday, January 10.

Support for researchers preparing for the extended storage and O2 outage is available through rchelp@hms.harvard.edu.

Research.files will be unavailable Friday, January 17, 2025, from 9:00 AM to 5:00 PM EST (UTC-5).

After the major migrations are completed, additional smaller moves will be scheduled individually with the affected groups. These moves are designed to minimize disruption and ensure a smooth transition.

These additional moves are expected to be completed by January 31.

Ensure that any essential tasks are scheduled outside of the planned downtime. Avoid relying on affected systems during the service disruption.

For example:

• Schedule O2 jobs in advance so they will be completed before January 3.
• If prompted, reset your HarvardKey or HMS account password before or after the January 3–4 outage.

If you must retain access to specific essential files and documents during the expected downtime, save a copy to unaffected storage systems like Harvard's Dropbox and Google Drive services, or to external hard drives. Ensure these copies are completed before the scheduled downtime.

Download any needed files or email before the outage begins on January 3. You may be able to remain connected to Microsoft 365 services if you are signed in on a phone or desktop client before the outage commences, but any interruption of service may keep you signed out until service is restored.

HMS IT will ensure the security and integrity of your data for this project, so you should only back up critical files that you need during the downtime.

Users who provide services to the community that will be impacted by this outage should post an outage notification to customers so they can also plan ahead.

Proactively inform your team or collaborators about the scheduled downtime. Share updates regarding unavailable systems. Coordinate deadlines and deliverables around the impacted period. Encourage team members to prepare their own data and workflows accordingly.

Ahead of the update, standby storage will undergo a temporary change. The snapshot expiration period will be reduced from eight weeks to two weeks during November and December.

This adjustment will help accelerate the transition by limiting the amount of data retained over shorter periods. The change will remain in place until the relocation to the new system is complete.

Take extra care when deleting files during this time, as snapshots will no longer be recoverable after two weeks.

Regularly check for updates via email notifications sent to your inbox and by visiting this project page.

Adjust your plans based on updated timelines or unexpected delays.

If you need help preparing for the data center relocation, use the following resources:

• Research-related support – Email rchelp@hms.harvard.edu for assistance with research systems like OMERO, eLabNext, or O2 computing.
• General IT support – Contact the HMS IT Service Desk at itservicedesk@hms.harvard.edu or call 617-432-2000.

HMS webpages and related services will be affected by scheduled outages from January 3 to 10. The level of impact depends on the hosting platform and specific service.

• HMS flagship website, microsites, OpenScholar, and Harvard Web Publishing (HWP) sites – These may remain viewable during the January 3-4 outage, but you will not be able to sign in or make changes.
• Websites and applications hosted on O2 – These will experience downtime during the January 3-10 outage.
• URLs redirecting to built-in source platforms – Vanity HMS URLs that redirect to a platform URL will return errors during the January 3–4 outage.

If relevant, notify your website users about the impending outages.

If your website uses vanity HMS URLs, consider distributing URLs that point directly to the service to minimize disruption during the outage.

Access to software app services hosted in the cloud, such as Salesforce and ServiceNow, may be disrupted during the January 3–4 outage.

Any service redirected from an hms.harvard.edu URL will encounter errors during the outage. Use the direct URL provided by the vendor to bypass redirection issues.

Sign-in attempts that use your HMS account may fail during this time.

Recommendations:

• Confirm and bookmark the direct URLs for your apps in advance.
• Reach out to your vendors for direct  URLs or additional guidance.
• Plan for potential disruptions and adjust workflows accordingly.

Remote access to desktop computers and services hosted in the data center will be unavailable during the move. This includes VPN connections and remote desktop services.

Recommendations:

• Plan ahead to avoid reliance on remote access during the outage period.
• If critical files or applications are needed, ensure they are accessible offline or via other methods.
• Communicate with your team to adjust workflows and timelines as necessary.

After services are restored, some initial slowness is expected as systems come online and users sign in simultaneously. Performance will stabilize shortly after all systems are fully operational.

Be patient during the initial period following restoration. Report any prolonged performance issues to IT support for further assistance.

HMS IT has developed a contingency plan that ensures backup systems are ready to be deployed to minimize service disruptions in case of major issues during the move.

As we inch closer to the end of the calendar year, we want to remind you about an upcoming HMS IT service project that will take place in early January and will result in widespread service disruptions. This is necessary due to the relocation of our data center, which we first announced in September.

We want you to be aware of the following so you can plan accordingly:

• All services provided or facilitated by HMS IT will be disrupted from Friday, Jan. 3, at 6 p.m., until Saturday, Jan. 4, at noon EST.
• HMS and HSDM building access, basic internet access, and HUIT-provided services that leverage HarvardKey authentication will not be affected during this time.
• Disruptions to O2 computing services and attached storage will continue through Friday, Jan. 10. This impacts researchers only and, therefore, education and administrative units should not be impacted. Support for researchers preparing for this outage is available by emailing rchelp@hms.harvard.edu.

While we have done our best to minimize the length and impact of these disruptions, we understand that any disruption impacts members of our community and your work. We appreciate your flexibility and understanding during this time.

This project represents an important investment in modernizing our IT infrastructure, enabling the stability and scalability required to support the innovative work happening across HMS and HSDM.

For more information about this project, including the timeline and impact, how to prepare, and answers to frequently asked questions, visit HMS IT’s Markley data center relocation webpage. We also encourage you to read this news story Q&A.

For all other questions, contact HMS IT by phone at 617-432-2000 or by email at itservicedesk@hms.harvard.edu.

To improve performance and keep our storage system updated, HMS IT will migrate research.files and admin.files data to a new storage system on Friday January 17, 2025, from 9:00 AM to 5:00 PM EST (UTC-5). 

During the migration, some files will be unavailable, including data on: 

• File servers – research.files, admin.files, and commons.med.harvard.edu 
• Web and Advanced Research Platforms (WARP) home directories – research.files.med.harvard.edu/home 
• O2 high-performance computing (HPC) transfer cluster - /n/files 

To help keep this outage as brief as possible, please avoid writing large numbers (e.g., millions) of files anywhere on the affected file systems mentioned above during the 24 hours before January 17th. Synchronizing data between the current and new storage servers takes time, so adding large data sets before the final sync will result in delays. 

Be sure to save any work stored in these affected directories before the migration begins. Do not try to access files in these locations during the migration. The files will be available again after the migration is complete. 

If you have any questions or concerns, contact Research Computing at rchelp@hms.harvard.edu.   

HMS IT is beginning a project in the Markley data center that will relocate existing services, consolidate servers, reduce power consumption, and decommission outdated hardware to improve efficiency, enhance resiliency, and lower costs. 

Benefits to the HMS community 

• Enhanced efficiency and increased sustainability: By optimizing and consolidating existing systems, we will improve performance, reduce energy consumption, and lessen our environmental impact. This will make operations more efficient and create room for future growth. 
• Improved infrastructure: Replacing aging infrastructure will enhance our capabilities and strengthen data center reliability for the long term. 
• Greater resiliency: With the addition of our new secondary data center at the Massachusetts Green High Performance Computing Center (MGHPCC), we’ve increased computational resiliency. MGHPCC is a state-of-the-art facility designed to support advanced research, innovation, and collaboration. 
• Cost savings: By optimizing services, improving power efficiency, and consolidating our data center footprint, we expect to reduce operational costs by 40% over the next 10 years. 

Impact 

The main work is currently planned for November 2024 through January 2025. During this period, there will be some scheduled downtime for key services including but not limited to O2, research.files, and admin.files, with the most significant outage expected in January. We will provide detailed dates and guidance to impacted users in advance to ensure they are fully informed and prepared. In addition, users can bookmark this web page which will be updated as information becomes available. 

During this move to a new data center, standby storage will be transitioning to a new system. To accommodate this system change, the snapshot expiration period for standby storage will be reduced from eight weeks to two weeks. This adjustment is intended to accelerate standby storage transition to the new system by limiting the amount of data retained over shorter periods until the relocation is complete. Researchers are encouraged to take special care when deleting files during this time period as they will not be recoverable after two weeks. 

Just as with the Research Data Migration project, HMS IT will strive to minimize the impact to the research community by scheduling downtimes during off-peak hours as much as possible, using backup systems, thoroughly testing infrastructure, maintaining clear communication, and coordinating closely with our vendors to ensure minimal disruption. 

Consultation and support  

Please reach out if you have questions about this project or its impact. Support for researchers is available through rchelp@hms.harvard.edu. For all other questions, you contact the HMS IT Service Desk at 617-432-2000 or itservicedesk@hms.harvard.edu.