The result is equivalent to replacing the target data directory with the source one. Only changed blocks from relation files are copied; all other files are copied in full, including configuration files. The advantage of pg_rewind over taking a new base backup, or tools like rsync, is that pg_rewind does not require reading through unchanged blocks in the cluster. This makes it a lot faster when the database is large and only a small fraction of blocks differ between the clusters.
pg_rewind examines the timeline histories of the source and target clusters to determine the point where they diverged, and expects to find WAL in the target cluster's pg_wal directory reaching all the way back to the point of divergence. The point of divergence can be found either on the target timeline, the source timeline, or their common ancestor. In the typical failover scenario where the target cluster was shut down soon after the divergence, this is not a problem, but if the target cluster ran for a long time after the divergence, the old WAL files might no longer be present. In that case, they can be manually copied from the WAL archive to the pg_wal directory, or fetched on startup by configuring recovery.conf. The use of pg_rewind is not limited to failover, e.g. a standby server can be promoted, run some write transactions, and then rewinded to become a standby again.
When the target server is started for the first time after running pg_rewind, it will go into recovery mode and replay all WAL generated in the source server after the point of divergence. If some of the WAL was no longer available in the source server when pg_rewind was run, and therefore could not be copied by the pg_rewind session, it must be made available when the target server is started. This can be done by creating a recovery.conf file in the target data directory with a suitable restore_command.
pg_rewind requires that the target server either has the wal_log_hints option enabled in postgresql.conf or data checksums enabled when the cluster was initialized with initdb. Neither of these are currently on by default. full_page_writes must also be set to on, but is enabled by default.
If pg_rewind fails while processing, then the data folder of the target is likely not in a state that can be recovered. In such a case, taking a new fresh backup is recommended.
pg_rewind will fail immediately if it finds files it cannot write directly to. This can happen for example when the source and the target server use the same file mapping for read-only SSL keys and certificates. If such files are present on the target server it is recommended to remove them before running pg_rewind. After doing the rewind, some of those files may have been copied from the source, in which case it may be necessary to remove the data copied and restore back the set of links used before the rewind.
pg_rewind accepts the following command-line arguments:
When --source-server option is used, pg_rewind also uses the environment variables supported by libpq (see Section 34.14).
When executing pg_rewind using an online cluster as source, a role having sufficient permissions to execute the functions used by pg_rewind on the source cluster can be used instead of a superuser. Here is how to create such a role, named rewind_user here:
CREATE USER rewind_user LOGIN; GRANT EXECUTE ON function pg_catalog.pg_ls_dir(text, boolean, boolean) TO rewind_user; GRANT EXECUTE ON function pg_catalog.pg_stat_file(text, boolean) TO rewind_user; GRANT EXECUTE ON function pg_catalog.pg_read_binary_file(text) TO rewind_user; GRANT EXECUTE ON function pg_catalog.pg_read_binary_file(text, bigint, bigint, boolean) TO rewind_user;
When executing pg_rewind using an online cluster as source which has been recently promoted, it is necessary to execute a CHECKPOINT after promotion so as its control file reflects up-to-date timeline information, which is used by pg_rewind to check if the target cluster can be rewound using the designated source cluster.
The basic idea is to copy all file system-level changes from the source cluster to the target cluster: