To be used when calibrations produced by several EPNs have to be sent to a single node, the aggregator. On the aggregator, the devices producing calibration will run and send the output to the CCDB.
- In order to populate the CCDB, the CCDB local server should be started, if the exercise is not meant to upload the official or test CCDB. In a terminal, run:
java -jar local.jarwhich will start a CCDB server on port 8080. See instructions.
- To run the calibration and aggregator, open a terminal and start the aggregator, which is just the
o2-dpl-raw-proxy:
o2-dpl-raw-proxy --dataspec A:TOF/CALIBDATA/0 --channel-config "name=readout-proxy,type=pull,method=bind,address=tcp://localhost:30453,rateLogging=1,transport=zeromq"where, as you can see, you can overwrite the configurations of the readout-proxy channel, which is the channel used by the aggregator.
The aggregator should be started before the other devices, since it will listen for data (it is the channel that is binding).
- The following example will pass the data arriving to the aggregator to the
o2-calibration-lhc-clockphase-workflowdevice, and from there to theo2-calibration-ccdb-populator-workflowto update the CCDB. For this, the raw-proxy should be in pipeline with these workflows like this:
o2-dpl-raw-proxy --dataspec A:TOF/CALIBDATA/0 --channel-config "type=pull,method=bind,address=tcp://localhost:30453,rateLogging=1,transport=zeromq,name=readout-proxy" | o2-calibration-lhc-clockphase-workflow --tf-per-slot 20 -b | o2-calibration-ccdb-populator-workflow --ccdb-path localhost:8080- To send data to the aggregator simulating an arbitrary number of EPNs, in another terminal, run:
source runEPNsimulation.sh 3where the argument (3 above) is the number of EPNs to be simulated. This will send the data to the aggregator process.
N.B.: the aggregator and calibration devices will need to use a different port from localHost:8080 in case the local CCDB server is used in its default configuration, since 8080 is used by CCDB. In the example above, port 30453 is used.