Different countries in Europe still suffer of elevated arsenic (As) concentration in groundwaters used for human consumption. In the case of households not connected to the distribution system, decentralized water supply systems, such as Point of Use (POU) and Point of Entry (POE), offer a direct benefit for the consumers. Field scale ex-situ treatment systems based on metallic iron (ZVI) are already available for the production of reduced volumes of drinking water in remote areas (village scale). To address drinking water needs at larger scale, we designed a pilot unit able to produce an elevated daily volume of water for human consumption. We tested the long-term As removal efficiency of a two steps ZVI treatment unit for the production of 400 L/day clean water based on the combination of ZVI corrosion process with sedimentation and retention of freshly formed Fe precipitates. The system treated 100 μg/L As(V)-contaminated oxic groundwater in a discontinuous operation mode at a flow rate of 1 L/min for 31 days. Final removal was 77–96% and the most performing step was aeration/sedimentation (A/S) tank with a 60–94% efficiency. Arsenic in the outflow slightly exceeded the drinking water limit of 10 μg/L only after 6000 L treated and Fe concentration was always below 0.2 mg/L. Under proposed operating conditions ZVI passivation readily occurred and, as a consequence, Fe production sharply decreased. Arsenic mobility attached to particulate was 13–60% after ZVI column and 37–100% after A/S tank. Uniform amorphous cluster of Fe nanoparticles (100 nm) formed during aeration drove As removal process with an adsorption capacity corresponding to 20.5 mgAs/gFe. Research studies often focus only on chemico-physical aspects disregarding the importance of biological processes that may co-occur and interfere with ZVI corrosion, As removal and safe water production. We explored the microbial transport dynamics by flow cytometry, proved as a suitable tool to monitor the fate of both single cells and bioactive particles along the treatment train of the pilot unit. A net release of bioactive particles, representing on average 26.5% of flow cytometric events, was promoted by the ZVI filter, with densities 10 times higher than those found in the inflow. In conclusion, the proposed system was efficient to treat large daily volumes of As contaminated groundwater. However, filter design and operating conditions should be carefully adapted to specific situation, since several key factors affect As removal efficiency. An effort in the optimization of ZVI filter design should be made to reduce fast observed ZVI passivation and low As adsorption capacity of the whole filter. More attention to biomass retention and bioactive particles travelling within the unit should be given in order to elucidate bacteria influences on As removal efficiency and related sanitary risks on long term basis.