During upgrades of our application, we generally shutdown all Secondary servers which are getting stream replicated from Primary Servers. This is to maintain a copy of database on other servers should we wish to revert (of course we take DB Backups too before starting the activity). After the application upgrade is done, when we start the secondary, often the replication is broken, and we need to again setup using pg_basebackup. How do we ensure that secondary is able to resume the replication without the need of base back up again?
Below are some of the WAL related settings:
shared_buffers = 48GB
wal_level = replica
max_prepared_transactions = 200
max_wal_senders = 5
wal_keep_segments = 32
hot_standby = ON
effective_cache_size = 144GB
work_mem = 1GB
maintenance_work_mem = 2GB
wal_buffers = 16MB
min_wal_size = 1GB
max_wal_size = 2GB
Pranjal Shukla