⚠ As of May 21st 2024 Partners MGB Network Security has implemented restrictions for external / remote connectivity to all DFCI and MGB servers. This includes ada and noah which are no longer accessible without an internal network connection via on-prem wifi or off-site VPN.
Remote / off-site access to any of our servers will require VPN connectivity or use of additional software for non-affiliated collaborators who cannot use the Partners MGB VPN.
DFCI staff / affiliated MGB personnel: please refer to this site (click) to log in, download the software and find instructions on obtaining and using the institute's VPN. If you encounter any issues with this contact Systems via email or Slack.
If you or a colleague needs access to our servers and you are not a DFCI or MGB employee with VPN access, please contact systems@ds.dfci.harvard.edu and specify your needs so we can help you with the alternate solution for server access.
⚠ Login servers are used for general computing and for laterally SSHing to other hosts. Please use the Kraken HPC cluster for large and multi-core jobs.
For guidance on accessing our servers, please refer to our instructions on this page.
For specific instructions on using RStudio, please see this page.
As a rule of thumb, please do not perform network-intensive data transfers, downloads, or uploads to any server above with a 1Gbps network speed, as this can greatly throttle the server's responsiveness for other colleagues who may be using the same system for their own work.
You are assigned 2 or 3 different accounts as part of your onboarding. The username will be the same on all Data Science department accounts (Unix, Windows and Google), along with the initial password, but you will be required to change them upon first login.
NOTE: These passwords are distinct and changing the password on one account does not mean that the password on the other two accounts is changed. There are different ways to change each account's password. Your departmental username and password will be emailed to you separately.
Login servers, such as ada, should not be used for resource-intensive jobs / commands. Resources are not only memory and CPU, but also bandwidth that can be consumed quickly by large scale file transfers.
To avoid service disruptions for yourself and others, please use ada only to reach other specialized servers, such as lyra for file transfers or libra for visualization jobs.