sercanto/infra
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

build03: update postgresql settings

Browse source
This commit is contained in:
zowoq 2024-07-27 11:16:46 +10:00
parent b41699aa7a
commit b120a9ebf7
1 changed files with 17 additions and 11 deletions

28
hosts/build03/postgresql.nix
View file

@ -18,18 +18,24 @@
enableJIT = true; enableJIT = true;
# From https://pgconfigurator.cybertec.at/ # https://pgtune.leopard.in.ua/#/
# https://pgconfigurator.cybertec.at/
# https://github.com/NixOS/infra/blob/4b5dd4f974d3f707b64ad60793b8182e645631ed/build/haumea/postgresql.nix
settings = { settings = {
# Connectivity # Connectivity
max_connections = 500; max_connections = 2000;
superuser_reserved_connections = 3; superuser_reserved_connections = 3;
# https://vadosware.io/post/everything-ive-seen-on-optimizing-postgres-on-zfs-on-linux/#zfs-related-tunables-on-the-postgres-side
full_page_writes = "off";
# Memory Settings # Memory Settings
shared_buffers = "8 GB"; shared_buffers = "32 GB";
work_mem = "32 MB"; work_mem = "128 MB";
maintenance_work_mem = "420 MB"; maintenance_work_mem = "2 GB";
huge_pages = "off"; huge_pages = "off";
effective_cache_size = "22 GB"; effective_cache_size = "64 GB";
effective_io_concurrency = 100; # concurrent IO only really activated if OS supports posix_fadvise function effective_io_concurrency = 100; # concurrent IO only really activated if OS supports posix_fadvise function
random_page_cost = 1.25; # speed of random disk access relative to sequential access (1.0) random_page_cost = 1.25; # speed of random disk access relative to sequential access (1.0)
@ -48,7 +54,7 @@
checkpoint_completion_target = 0.9; checkpoint_completion_target = 0.9;
# 2x default, hint from service logs # 2x default, hint from service logs
max_wal_size = "2 GB"; max_wal_size = "5 GB";
min_wal_size = "1 GB"; min_wal_size = "1 GB";
# WAL writing # WAL writing
@ -64,10 +70,10 @@
bgwriter_flush_after = 0; bgwriter_flush_after = 0;
# Parallel queries: # Parallel queries:
max_worker_processes = 8; max_worker_processes = 24;
max_parallel_workers_per_gather = 4; max_parallel_workers_per_gather = 12;
max_parallel_maintenance_workers = 4; max_parallel_maintenance_workers = 12;
max_parallel_workers = 8; max_parallel_workers = 24;
parallel_leader_participation = "on"; parallel_leader_participation = "on";
# Advanced features # Advanced features