Skip to main content
Use case study of Hive/HadoopEva Tse,Jerome Boulon
What are we trying to achieve?Scalable log analysis to gain business insights:Logs for website streaming (phase 1)All logs from web (phase 2)Output required:Engineers access:Ad-hoc query and reportingBI access:Flat files to be loaded into BI system for cross-functional reporting.
Some Metrics	Parsing 0.6 TB logs per dayRunning 50+ persistent nodes
Architecture Overview Web AppPhase 1Phase 2Phase 2Chukwa CollectorLog copy deamonHive & Hadoop (for query)Hive MetaStore     		S3					  HDFS   /     S3Hive & Hadoop running on the cloud
Chukwa StreamingMyAppCollectorNo data written to disk on the application side
Data sent to a remote collector using Thrift
Collector write to localFS/S3n/HDFS compressed
http://wiki.github.com/jboulon/Honu/ (stay tuned)Workflow to Hive (phase 1)Streaming Session reconstructionEach hour: Run hadoop job to parse last hour log and reconstruct sessions merge small files (from each reducer) load to hiveSession expiration after 24 hours.Will have sessions for each of past 24 hours.After 24 hours, will need to merge again by: insert overwrite …. select <column list> from table
Workflow to Hive (phase 2)Continuous log collection via ChukwaGeneric and continuous parse/merge/load to ‘real-time’ Hive warehousemerge at hourly boundary and load to public Hive warehouse. SLA is 2 Hr on merged data.Daily/Hourly job: For summary.For publishing data to BI for reporting.