Apache HBase: crazy dances on the elephant back
summary
HBase runs as a real‑time, column‑family store on top of Hadoop HDFS/YARN with ZooKeeper coordinating masters and region servers; writes go to WAL + memstore append for fast ack and later compaction to HFiles, replication defaults to three, and large clusters (thousands of nodes) expect daily node failures. Typical workloads are single‑row gets and range scans; production integration includes bulk-loading HFiles from MapReduce, pushing updates into Solr via an indexer, and using region coprocessors for distributed hash/aggregation — powerful but operationally tricky.
AI-written summary · about
chapters (10)
- 0:00Introduction to Apache HBase
- 1:15What is Apache HBase
- 6:42Hadoop and HBase
- 10:19HBase data model
- 17:05Physical data structure
- 22:28Read and write operations
- 28:11Server-side scanning
- 33:57MapReduce integration
- 37:11Secondary indexes and search
- 43:18Q&A
related talks
32mContaining an Elephant
53mBig Data: from mammoth to elephant
20mReal-time Analytics with HBase
51mBIG DATA STORAGE & PROCESSING: Introduction to Hadoop and Its Ecosystem
59mHadoop Essentials for the SQL Server Professional
29mMulti Tenancy — Apache Hbase
01h 29mHadoop 201 -- Deeper into the Elephant
43mMulti-tenant HBase Solutions at Facebook