OpenFOAM Cloud Backend
Installing from GitHub
Clone the repository
git clone https://github.com/mikelangelo-project/openfoam-cloud.git
Install required Python packages
pip install -r requirements.txt
Installing via pip
pip install openfoam-cloud
Install the following required packages
Copy local settings
Edit local settings to reflect your environment. Where you see os.environ['ENV_VARIABLE_NAME'] you have to export the desired value to your environment.
OPENFOAM_BASENAME = 'http://openfoam.example.com' GRAFANA_BASENAME = 'http://grafana.example.com' S3_ACCESS_KEY_ID = os.environ['S3_ACCESS_KEY_ID'] S3_SECRET_ACCESS_KEY = os.environ['S3_SECRET_ACCESS_KEY'] S3_HOST = 's3-host' S3_PORT = 8443 SNAP_SERVICE = 'snap-api-root' INFLUX_DB_HOST = 'influxdb-host' INFLUX_DB_PORT = 8086 INFLUX_DB_NAME = 'snap' INFLUX_DB_USER = 'root' INFLUX_DB_PASS = 'root' # Dedicated openfoam network settings OPENFOAM_NETWORK_PREFIX = 'openfoam' OPENFOAM_NETWORK_CIDR = 'your desired network CIDR' OPENFOAM_NETWORK_ALLOCATION_POOL_START = 'your IP allocation range start' OPENFOAM_NETWORK_ALLOCATION_POOL_END = 'your IP allocation range end' OPENFOAM_NETWORK_GATEWAY_IP = 'your gateway IP usually the first IP in CIDR range' # Scheduler daemon settings SCHEDULER_REFRESH_INTERVAL_SECONDS = 'desired scheduler daemon refresh interval in seconds' # OpenFOAM simulations save their results on a NFS server as is evident from the NFS_IP setting. The # LOCAL_NFS_MOUNT_LOCATION setting tells the scheduler daemon where to prepare simulation case files, capstan package etc. # This folder should have the NFS location mounted (example /mnt/OpenFOAM_results) except when the scheduler runs on # the NFS machine itself, then this can point directly to the exported directory (for instance '/export/OpenFOAM_results/') LOCAL_NFS_MOUNT_LOCATION = 'path to folder with mounted NFS location' # Network file storage server ip address NFS_IP = 'nfs server ip' # Location on the NFS server where OpenFOAM case files and results will be saved NFS_SERVER_MOUNT_FOLDER = 'location on the nfs server, where simulation case files and results are saved' # Maximum number of launch retries of one instance. When this limit is reached, the simulation instance enters the # 'FAILED' state OPENFOAM_SIMULATION_MAX_RETRIES = 3 # Overrides the number of maximum NOVA vcpu's used. Example: openstack allows use of maximum 24 vcpus. Our setting # allows only 12. The scheduler_deamon will run new simulations until 12 vcpu's are used on nova. # If this setting value is higher than nova's max VCPU quota, the latter will be respected. OPENFOAM_MAX_CPU_USAGE = 18 # Overrides the number of maximum NOVA instances used. Example: openstack allows use of 10 instances. Our setting # allows only 5. The scheduler_deamon will run new simulations until 5 oinstances are used on nova. # If this setting value is higher than nova's max instance quota, the latter will be respected. OPENFOAM_MAX_INSTANCE_USAGE = 8
Ensure snap, influxdb and grafana are running. Define a datasource for influxdb in grafana.
/opt/snap/bin/snapd -t 0 -a /opt/snap/plugin
Then launch the OpenFOAM backend
python manage.py runserver 0.0.0.0:8008
Next launch the OpenFOAM scheduler daemon
python manage.py runscheduler
This project has been conducted within the RIA MIKELANGELO project (no. 645402), started in January 2015, and co-funded by the European Commission under the H2020-ICT- 07-2014: Advanced Cloud Infrastructures and Services programme.