Transaction log archiving and recovery
After this brief introduction to the transaction log in general, it is time to focus our attention on the process of transaction log archiving. As you have seen, the transaction log contains a sequence of binary changes made to the storage system. So, why not use it to replicate database instances and do a lot of other cool stuff?
Configuring for archiving
The first thing we want to achieve in this chapter is to create a configuration to perform standard Point-In-Time Recovery (PITR). There are a couple of advantages of PITR over ordinary dumps:
- You will lose less data because you can restore to a certain point in time and not just to the fixed backup point
- Restoring will be faster because indexes don't have to be created from scratch. They are just copied over and are ready to use
Configuring for PITR is easy. Just a handful of changes have to be made in the postgresql.conf file:
wal_level = replica # used to be "hot_standby" in older versions max_wal_senders...
