Application Information

Architecture

HTCondor is installed as a service on the Rubin development nodes (rubin-devl) with Ansible. The HTCondor central collector service is installed on sdfiana12. If there is a failure on sdfiana12 then the central collector is installed on sdfiana13.

HTCondor interfaces with Slurm and creates a Condor Pool on the Slurm nodes.

Architecture Diagram

Associated Systems

ctrl_bps_htcondor is a HTCondor pluguin for LSST PipelineTask execution. HTCondor interacts with Slurm for job execution.

Configuration Location

Config Area	Location
Configuration
Vault Secrets Dev	None
Vault Secrets Prod	None

Data Flow

Dependencies - S3DF

• Weka Filesystem for data for jobs. HTCondor can run without Weka, but would have no data for jobs.

• Slurm

• Rubin Development Nodes. sdfiana12 is most important because it runs the HTCondor central collector service

• CVMFS Software stack

Dependencies - External

No external dependencies. HTCondor is currently not integrated with multi-site processing.

Disaster Recovery