Production to merge histograms are slightly different from normal data processing ones. There are two different type of input files to consider: BRUNELHIST and DAVINCIHIST, so in the Input Data section of the request, it must be selected both file types.

• Screenshot of the Histo Merging request:
  

The template to choose is everyThingElse_run.py. Type HistoMerge as production to be launched, GroupSize must be equal to 999999, as we create one single merged file per run. Output Data Storage Element must be CERN-HIST-EOS. The plugin name must be ByRunSizeWithFlushRootMerging and the removeInputs flag must be set to True in order to delete the unmerged files.

-- MarcoCorvo - 2016-10-19