Analytics WG meeting

Wednesday 25 Mar 2015, 14:00 → 16:00 Europe/Zurich

31/S-023 (CERN)

News:

• Hardware specs from procurement and Ceilometer data should be available in the near future
• DB and DSS group will run a Hadoop service together. DSS will provide infrastructure and DB will concentrate on higher level software (Impala etc.)
  • There wil first be a prototype service
  • Initial discussions with Cloudera regarding official support license are underway
  • Some SDC applications might need production level support in summer, meaning there is a demand for a full fledged service

Hadoop Job Monitoring:

• We need to figure out, how to monitor and control competition between several users/jobs on the hadoop cluster

14:00 → 14:05 Minutes & News

Speaker: Dirk Duellmann (CERN)

14:05 → 14:25 Monitoring of Job Execution in Hadoop

Speaker: Luca Menichetti (CERN)

Slides awg-meeting-20150325.pdf awg-meeting-20150325.pptx

14:25 → 14:45 Hadoop vs. Spark - Update on performance comparison

Speakers: Luca Menichetti (CERN), Vag Motesnitsalis (Imperial College Sci., Tech. & Med. (GB))