I’m new to the postgresql. Come across this issue and has been scratching my head for couple of days now. Can someoen explain below behavior. I have 2 postgresql-10 servers, one master one slave. Master has following config
wal_level = replica max_wal_senders = 10 wal_keep_segments = 10 ... Replication works when both servers are online.
However, when slave server gone offline (server crash or network issue), I found the master server won’t be able to create or drop database.
postgres=# create database abc; The console just stuck there.
ps output gives
postgres 1442 977 0 13:08 ? 00:00:00 postgres: postgres postgres [local] CREATE DATABASE waiting for 0/A004FA8 From what I understand, it means the master can’t write WAL to file? Probably due to loss of connection to slave. So this begs the question that replication creates a failure point? if for whatever reason slave server is unavailable, master server won’t be able to server frontend properly? is this against the conception of resilience where any node failed, the other one can restore the service quickly even if manual intervention is required?
