This note documents the mechanisms of logging and recovery.

1. Logging: As transactions normally proceed, some records are made so that in case of a crash, these records can be used for recovery.
2. Recovery: If a transaction crashes/fails, after the system restarts, the database needs to perform recovery operations.

The previous lesson introduced Two-Phase Locking Concurrency Control, a method of concurrency control implemented using the lock mechanism. This note focuses on Timestamp Ordering Concurrency Control, which is based entirely on timestamps and does not use locks. In this lesson, you can understand:

• When Timestamp is distributed, to whom and how it is distributed
• What Timestamp Ordering Concurrency Control is
• What is the difference between strict T/O
• Isolation Level

This lesson mainly introduces the Two-Phase Locking mechanism. After learning, you can master:

• What is Two-Phase Locking
• What is Rigorous Two-Phase Locking, why rigorous is needed
• How to handle deadlock
• Why and how to use Intention lock

This lesson introduces the Concurrency Control and Recovery section from a high level perspective, and the specific implementation will be introduced in the next lesson. Through this lesson we can understand:

• What is Transaction, Schedule
• Why we need Concurrency Control and Recovery
• ACID corresponding to the case
• Serial Schedule, Serializable Schedule, Conflict Serializable Schedule
• How to determine Conflict Serializable Schedule

This article discusses the difference between on-policy and off-policy methods in reinforcement learning. On-policy methods use data derived from the same policy to optimize the target policy, while off-policy methods use data from a behavior policy for optimization. The concepts are exemplified with SARSA and Q-learning.