Class RegionReplicaFlushHandler

java.lang.Object
org.apache.hadoop.hbase.executor.EventHandler
org.apache.hadoop.hbase.regionserver.handler.RegionReplicaFlushHandler
All Implemented Interfaces:
Comparable<EventHandler>, Runnable

@Private public class RegionReplicaFlushHandler extends EventHandler
HBASE-11580: With the async wal approach (HBASE-11568), the edits are not persisted to WAL in secondary region replicas. This means that a secondary region replica can serve some edits from it's memstore that are still not flushed from primary. We do not want to allow secondary region's seqId to go back in time, when this secondary region is opened elsewhere after a crash or region move. We will trigger a flush cache in the primary region replica and wait for observing a complete flush cycle before marking the region readsEnabled. This handler does the flushing of the primary region replica and ensures that regular region opening is not blocked while the secondary replica is blocked on flush.