Stolon can use pg_rewind to speedup instance resynchronization (for example resyncing an old master or a slave ahead of the current master) without the need to copy all the new master data.
It can be enabled setting to true the cluster specification option usePgrewind (defaults to false):
stolonctl [cluster options] update --patch '{ "usePgrewind" : true }'This will also enable the wal_log_hints postgresql parameter. If previously wal_log_hints wasn't enabled you should restart the postgresql instances (you can do so restarting the stolon-keeper)
pg_rewind needs to connect to the master database with a superuser role (see the Stolon Architecture and Requirements).