High Availability and Recovery
What keeps a cluster alive when a host dies, and what happens in the seconds after it does.
-
Chapter 20.1
Write-Ahead Logging in Apache Cloudberry
WAL records, the resource managers Cloudberry adds, checkpointing, and how distributed commit is encoded so that a crash can be replayed correctly.
-
Chapter 20.3
Streaming Replication, Recovery and FTS in Cloudberry
Segment mirroring over streaming replication, crash and in-doubt 2PC recovery, and how the fault tolerance service detects and fails over a segment.