As an in-memory computing platform, GridGain and Apache Ignite support native persistence that stores data and indexes transparently on non-volatile memory, SSD or disk. When persistence is enabled, memory becomes a cache for the most frequently used data and indexes. Native persistence is ACID-compliant, durable and enables immediate availability on a restart of each node. Data is never lost; GridGain supports full and incremental snapshots along with continuous archiving, and provides Point-in-Time recovery to an individual transaction.
Join Ivan Rakov, one of the senior engineers who developed native persistence, backup, and Point-in-Time recovery as he provides his insights into the underlying architecture and best practices for implementing native persistence in production. This session will cover:
- An architectural overview of native persistence, and centralized backup and recovery
- Tips and tricks for configuring and managing persistence
- Best practices for checkpointing, using the Write-Ahead Log (WAL) and restoring from a failure
- Performance tuning recommendations to balance durability and performance, including how to create snapshots under load