Managing WAL files with walctl
The walctl toolkit provides two extra scripts that a DBA should never have to call manually: walctl_push and walctl_pull. These are purely intended to facilitate the preferred architecture of walctl. However, we also understand that many PostgreSQL servers exist already, and not every cluster is new.
It's actually very likely that at least one clone exists now and that such behavior is directly supported by PostgreSQL 9.1 and more. In this recipe, we'll explore how to convert an existing cluster to use walctl for WAL management instead.
Getting ready
This recipe depends on walctl being installed on a primary server and any existing PostgreSQL clones. Please follow the Installing and configuring walctl recipe before continuing.
How to do it...
For this recipe, imagine we have four PostgreSQL servers. The primary PostgreSQL server is pg-primary, and we also have three existing replicas named pg-clone1, pg-clone2, and pg-clone3. Execute this command as the postgres...
