EventServer for data export

Currently there are two ways built into mvme which allow readout data to be accessed by external software:

• The Analysis ExportSink which is described in the Analysis reference.

  This method should be used when a limited number of analysis data arrays - all belonging to modules in the same VME event - are to be exported.

• The EventServer component described in this section. It provides export functionality of the complete event data produced during a DAQ run. The data is transmitted over a TCP network connection.

  The EventServer should be used when you want to export all of the data produced by an experiment.

  The source code for a minimal example client and for a ROOT client is shipped with mvme.

Overview

The EventServer component is a TCP based network server built into mvme. It uses a custom protocol to transmit extracted VME module data across the network.

The “extracted” data is produced by the data filters attached to each of the VME modules present in the experiment setup. E.g. with the default filters for a MTDC-32 module the server will send 32 time values, 2 trigger_times and a module timestamp.

Fig. 38 MTDC-32 with default data extraction filters.

The server is attached to the analysis side of mvme which means during a live DAQ run it will not slow down the readout. Instead if the server, or the attached clients, cannot keep up with the readout data rate then the normal buffer loss mechanism at the core of the mvme DAQ will start to discard data. This means during a live run clients may only see parts of the data, whereas when replaying from a mvme listmode file all data will be transmitted.

The supplied ROOT client recreates the full VME Event -> VME Module -> Data Array structure present in an mvme setup using automatically generated ROOT classes. Additionally histograms of the raw parameter values are created and written to a separate ROOT file.

Fig. 39 ROOT tree structure created from an MTDC-32.

Note

The ROOT client requires ROOT6 and has so far only been tested on Linux.

Note

An example mvme workspace which contains pre-recorded listmode data can be found on our website: https://mesytec.com/kundendaten/mvme-examples/mvme-example-workspace-01.tar.bz2

This workspace can be used to test the server and client without needing to record your own data first.

Using the EventServer

• In the mvme main window go to Workspace Settings and check the Enable Event Server box. Optionally enter an IPv4 address or a hostname to bind the server to a specific network interface.

  Also enter an Experiment Name in the box at the top of the dialog. This will affect the filenames produced by the mvme_root_client.

• For a first test the mvme_event_server_example_client can be used. It can be found in the bin/ directory of the mvme installation. Additionally the source code and a small Makefile are present in share/mvme_event_server_example_client.

  Start the client on the same machine: ./bin/mvme_event_server_example_client --print-data.

  The client should be able to connect to the server running on localhost.

• Open an example listfile via File -> Open Listfile and also load the analysis from the file so that some data extraction filters are present.

• Start the replay. The client prints the incoming data to stdout.

If you want to write your own client you can find the protocol definition and implementation of the base Client class in ${MVME}/include/mvme/event_server/common.

Using the ROOT client

The ROOT client is not shipped in binary form but has to be compiled manually so that it links against the locally installed version of ROOT.

To build the client run the following shell commands:

cd mvme                                        # cd into the mvme installation directory
source bin/initMVME                            # setup the enviroment ($PATH, $LD_LIBRARY_PATH)
make -C ${MVME}/share/mvme_root_client install # compile and install the client

You can now create a data directory to hold the generated code and output ROOT files and start the mvme_root_client from within this directory.

The mvme_root_client works in two modes:

1. In client mode it connects to an mvme instance and receives incoming readout data to produce ROOT output files.
2. The program replays data from a previously created ROOT output file.

In the first case the data description sent by mvme at the beginning of a run is used to generate ROOT classes. These classes are then compiled and loaded into the client. Compilation is done using a simple Makefile. The resulting ROOT library is then loaded via TSystem::Load().

During the run incoming event data is interpreted and used to fill instances of the generated ROOT classes. At the end of the run the class hierarchy is written to file in the form of TTrees. Additionally raw histograms are created, filled and written to a separate output file.

In the 2nd case the given input file is opened and mvme specific information is used to locate the previously built ROOT object library. The library is then loaded like in case 1) and the ROOT objects are filled from the data in the input file.

In both cases user-editable analysis code (analysis.cxx) is loaded (via dlopen()/dlsym()) and is invoked for each complete event received from mvme or read from the input file. For each defined VME event the client will attempt to call a function called analyze_<eventname>(), e.g. analyze_event0().

For details on what the generated classes contain see the <ExperimentName>_mvme.{h,cxx} files. The base classes are defined in ${MVME}/share/mvme_root_client/mvme_root_event_objects.{h,cxx}.