On 14 Listopad 2013, 16:09, Alexander Farber wrote: > pgtune has produced the following for my server (the specs: > http://www.hetzner.de/en/hosting/produkte_rootserver/px60ssd ): > > default_statistics_target = 50 > maintenance_work_mem = 1GB > constraint_exclusion = on > checkpoint_completion_target = 0.9 > effective_cache_size = 22GB > work_mem = 192MB > wal_buffers = 8MB > checkpoint_segments = 16 > shared_buffers = 7680MB > max_connections = 80 > > Is it really okay? Isn't 22GB too high? effective_cache_size is a hint on how much memory is there for filesystem cache, so that it can be considered when planning queries. PostgreSQL relies on filesystem cache, so this is needed to answer questions like 'how probable it's the block is in cache and won't actually cause any I/O"? It does not allocate anything. You do have 32GB of RAM in total, so 22GB for caches seems about right unless you're running other memory-intensive applications on the same box (making less memory to be available for the filesystem cache). Tomas -- Sent via pgsql-general mailing list (pgsql-general@xxxxxxxxxxxxxx) To make changes to your subscription: http://www.postgresql.org/mailpref/pgsql-general
