Breakout room #2: Improvement of analysis job efficienty at Belle II distributed computing system
Presented by Ms. Hikari HIRATA on 17 Nov 2020 from 14:20 to 16:00
Session: Virtual Poster Session
In the Belle II experiment, a distributed computing system is utilized for not only saving and processing huge data but also analyzing it for the physics. However, around 10% of analysis jobs submitted into the system were failed in 2019, due to errors in analysis scripts. Such failed jobs make computing resources waste, and they have potential to trigger the system troubles. In order to suppress failed analysis jobs, we developed a new framework. In the framework, syntax in the scripts is checked every before submitting analysis jobs. Moreover, when huge analysis jobs are submitted, a few test jobs are submitted on the system for the sanity check. In this poster, the details and the development status will be given.