- hello_server.py: A file to easily detect whether a port is available for remote connection.
- query_job_nodelist.bash: Helper script to locate a Slurm job to the execution hosts.
- workflow/: Using Cylc or Fireworks workflows to automate demos on JLab's
ifarm
cluster.
Demos for a 3-node system:
- Sender node (allocated with
sbatch
); - Receiver node (allocated with
sbatch
); - Control node: Centrailized metrics DB with Prometheus or InfluxDB to moniter the sender and rceiver node. If it's a compute node, it is allocated with the interactive mode.
May need minor updates (such as path) to run the scripts on ifarm.
-
slurm-iperf3-prom-demo/: The centralized metric DB is Prometheus. Process-exporters run on the sender (native
iperf3
client) and receiver (nativeiperf3
server) node. -
slurm-podio2tcp-influxdb-demo/: The centralized metric DB is InfluxDB. The sender node runs
podio2tcp
. The receiver node runstcp2podio
.
- sifs/: The Singularity/Apptainer containers and the method to pull them.
- config/ Configuration files or configuration helper files for Prometheus exporters, Prometheus, etc.