MemTable Crash-Safe Recovery

In our benchmark, recovering a 1.5 GiB MemTable after a process crash took an average of 4.989 ms with CSPP crash-safe recovery versus 5987.152 ms with default SkipList/WAL replay—about 1200× faster. Process crash: abnormal process termination, including deliberate termination with kill -9 / abort; the operating system continues running, and its file page cache (PageCache) remains intact. Hardware power loss: the server hardware suddenly loses power, the operating system stops running, and in-memory state (including PageCache) is lost. The container orchestration system Kubernetes allows a def
This is an AI-generated summary. ShortSingh links to the original source for the complete article.
Discussion (0)
Log in to join the discussion and vote.
Log in