Hardware: Computing Cluster

GM/CA @ APS computing environment has similar structure for all three beamlines: 23-ID-D, 23-ID-B and 23-BM. The 23-ID-D and 23-ID-B beamlines which are equipped with fast Eiger2-16M CdTe and Eiger-16M detectors respectively share two storage arrays: an ultrafast 150 TB buffer storage based on NVMe flash drives and 1.5 PB main storage consisting of 216 x 8 TB hard drives. Both storage arrays utilize BeeGFS distributed file system. The BM beamline equipped with the Pilatus3-6M detector provides a 0.5 PB shared storage with BeeGFS file system. The computing infrastructure of the ID beamlines clusters is shown on the picture below.

Users who log on the beamline workstations have their home directories on the BeeGFS storage array and thus all the workstations access the same home directory for a given user account. The storage capacity allows us to keep users data for three months after the experiment.

Computers at the ID beamlines are connected to internal 100 Gbps fiber network and the workstations accessible from outside the lab (ws2, ws5, ws6, and ws8) are connected to 100 Gbps uplink. The BM subnet has 40Gbps fiber network and 10 Gbps uplink.

Users are provided with two groups of workstations. One group is allocated for collecting and processing data on the day of experiment (day-1 workstations). These are blXws2, blXws6, and blXkeithley (inside end station), where "X" stands for the beamline number ('1' for IDD, '2' for IDB, and '3' for BM). The other group (day-2 workstations) consisting of ws5 and ws8 is offered for those users who wish to continue processing or backing up their data after the experiment is over. The other workstations ws9 to ws12 of IDD combined with ws9 to ws12 of IDB) are used for automatic data processing from BluIce. They do not have monitors and keyboards for user access, but can be ssh-ed to.

All computers operate under Linux Mint, a refined freeware clone of Ubuntu LTS Linux operating system, and MATE graphical desktop environment. All of them have most common crystallographic data processing software packages installed including HKL3000/HKL2000, Phenix, PyMol, and etc.

The following computing policies are implemented:

• The account management is centralized and all workstations access the same home directories that actually reside on the storage array.
• Users login with their ANL accounts during the provided beamtime and one day after. They will share a home directory allocated on the storage array for the experiment.
• No disk quota is enforced on user account.
• All workstations are provided with USB3 connectivity (which accepts USB2 as well, but the later is not recommended because of low speeds). Users are encouraged to bring their external drives for making data backups. More information about backups to external drives is provided on the data management webpage.
• Users can remotely download collected data to their institutions using the GMCA Globus servers.
• Users can also SFTP out their data. The speed rate may vary on the route to user's institution. The best expected rate is about 7-8MB/s. Due to tight ANL security restrictions, option to SFTP in, i.e. to access data from user's home institution is only available with ws5 for the day of experiment plus one day. Please inquire your host if you need such access.
• Users laptops using Wi-Fi can connect to ws2, ws5, and ws6 via SSH, SFTP or NOMACHINE protocols. The Wi-Fi connection also provides access to outside Internet resources like web pages and e-mail.
• GM/CA @ APS stores users data for three months from their experiment start date. During this period users are expected to verify that their backup was successful and data was safely delivered to their home institution. After this period an e-mail is sent to remind users about scheduled deleting of their data and then in two days the data is automatically deleted from the GM/CA storage array.
• User accounts are automatically disabled one day after their experiment start date. If you need to extend your access to day-2 workstations or to the Globus servers, please send request to your host who will arrange a temporary exception. Permanent exceptions are not possible.
• Remote access using NoMachine technology is possible. For additional details click here.