CMS Big Data Science Project
Description
FNAL room: Dark Side-WH6NW - Wilson Hall 2nd fl North East
CERN room: 513-R-068
Instructions to create a light-weight CERN account to join the meeting via Vidyo:
If not possible, people can join the meeting by the phone, call-in numbers are here:
The meeting id is hidden below the Videoconference Rooms link, but here it is again:
- 10502145
-
attendance: Vagg, David Lange, Thanat Jatuphattharachat, Lukasz, Saba, Siewan, Matteo, JimP, Igor, OLI, Jorge, Pratyush
-
ACAT
- spark-root
- new connector
- performance
-
presentations
- root4j/spark-root
- clean separation almost done
- Pratyush started classes, Victor will pick it up again
- GROOT test case, but would be usable to have root like plotting functionality from JAVA and scala
- Distributed query service
- Investigated: Apache Zookeeper, Mesos, Marathon
- Apache Zookeeper was investigated more closely
- Can build our workflow (FemtoCode) on top of Zookeeper
- root4j/spark-root
-
Discussion
- performance
- dedicated hadoop cluster (3 nodes, 2 data nodes)
- 10 Gbps network
- 3x 32 cores
- compare reading from HDFS on data nodes or from EOS over the network
- HDFS replication factor is 2, data is always local
- conclusion: connector does not add overhead
- question: walltime investigation, interesting, will be done later
- network performance
- tried to put ROOT files on CERNBOX, no, maybe just a test
- continue discussion offline
- dedicated hadoop cluster (3 nodes, 2 data nodes)
- performance
-
plan
- meet next week, August 16, to finalize ACAT content
There are minutes attached to this event.
Show them.