|
I have a setup with the following: One (1) "main" site Ten (10) "regional" sites "main" site is a publisher to each of the regional sites (for 20 tables). "main" site is also a subscriber from each of the 10 regional sites (5 tables). In short: The main site has 1 publication and 10 subscriptions. Each regional site has 1 publication and 1 subscription. This is in addition to the streaming replication that each
main and regional sites have for High Availability.
I currently have the following Postgres configuration for "main" site max_wal_senders (25) max_replication_slots (25) max_logical_replication_workers (25) max_worker_processes (25) logical_decoding_work_mem (128MB) max_slot_wal_keep_size 750MB max_parallel_workers (8) default max_parallel_maintenance_workers (2) default max_parallel_workers_per_gather (2) default max_sync_workers_per_subscription (2) default
I currently have the following Postgres configuration for "regional" site max_wal_senders (10) max_replication_slots (10) max_logical_replication_workers (10) max_worker_processes (10) logical_decoding_work_mem (128MB) max_slot_wal_keep_size 750MB max_parallel_workers (8) default max_parallel_maintenance_workers (2) default max_parallel_workers_per_gather (2) default max_sync_workers_per_subscription (2) default
Is this configuration sufficient and complete for a setup described above?
If I set higher number but do not use them (not all 10 regional sites defined) will it still consume resources? I prefer to set higher numbers now so I will not need to restart Postgres when more "regional" setup are installed as long as it is not consuming to much resources if currently not
used. Thanks! |
