Event reconstruction in the LHC$b$ Online cluster

The LHC$b$ experiment at the LHC accelerator at CERN will collide particle bunches at 40 MHz. After a first level of hardware trigger with output at 1 MHz, the physically interesting collisions will be selected by running dedicated trigger algorithms, the High Level Trigger (HLT), in the Online computing farm. This farm consists of 16000 CPU cores and 40 TB of storage space. Although limited by environmental constraints, the computing power is equivalent to that provided by all Tier-1's to LHC$b$. The HLT duty cycle follows the LHC collisions, thus it has several months of winter shutdown, as well as several shorter machine and experiment downtime periods. This work describes the strategy for using these idle resources for data reconstruction. Due to the specific features of the Online Farm, typical processing $à$ $la$ Tier-1 (1 file per core) is not feasible. A radically different approach has been chosen, based on parallel processing the data in farm slices of $O$(1000) cores. Single events are read from the input files, distributed to the cluster and merged back into files once they have been processed. A detailed description of this architectural solution, the obtained performance and how it will be connected to the LHC$b$ production system will be described.