In this chapter, you have already seen that setting up replication is easy and does not require a lot of effort. However, there are always some corner cases which can cause operational challenges. One of those corner cases is all about transaction log retention.
Consider the following scenario:
- A base backup is fetched
- After the backup, nothing happens for one hour
- The slave is started
Keep in mind that the master does not care too much about the existence of the slave. Therefore, the transaction log needed for the slave to start up might not exist on the master anymore as it might have been removed by checkpoints already. The problem is that a re- sync is needed to be able to fire up the slave. In the case ...
