[HBASE-451] Remove HTableDescriptor from HRegionInfo

• Type: Improvement
• Status: Closed
• Priority: Critical
• Resolution: Fixed
• Affects Version/s: 0.2.0
• Fix Version/s: 0.92.0
• Component/s: master, regionserver
• Hadoop Flags: Incompatible change, Reviewed
• Release Note: Removes HTableDescriptor from HRegionInfo. Runs a migration of catalog tables on startup.
• Assignee: Subbu M Iyer
• Reporter: Jim Kellerman
• Created: 15/Feb/08 01:16
• Updated: 03/Oct/20 22:41
• Resolved: 14/Jul/11 22:55

Description

There is an HRegionInfo for every region in HBase. Currently HRegionInfo also contains the HTableDescriptor (the schema). That means we store the schema n times where n is the number of regions in the table.

Additionally, for every region of the same table that the region server has open, there is a copy of the schema. Thus it is stored in memory once for each open region.

If HRegionInfo merely contained the table name the HTableDescriptor could be stored in a separate file and easily found.

Issue Links

• blocks: HBASE-922 Bloom filters will not work if enabled after table is created and contains data
• blocks: HBASE-3970 Address HMaster crash/failure half way through meta migration
• blocks: HBASE-484 Store HRegionInfo in META and ROOT as multiple cells instead of a binary field
• is blocked by: HBASE-546 Use Zookeeper in HBase
• is duplicated by: HBASE-565 Move Table Schema out of HRegionInfo
• is duplicated by: HBASE-4055 Client region location caches redundant HTableDescriptor's
• is part of: HBASE-1816 Master rewrite
• is related to: HBASE-4032 HBASE-451 improperly breaks public API HRegionInfo#getTableDesc
• is related to: HBASE-5204 Backward compatibility fixes for 0.92