I'm playing around with the effective cache size setting today. Someone recommended the formula
((buffers + cached ) / 8 ) during heavy querying to optimize the 'Postgresql.conf' setting.
I have 4 separate physical servers and I have all of them set to 512 MB, even though the results of the formula show.
Server 1 1 GB
Server 2 1 GB
Server 3 2 GB
Server 4 2 GB
These results indicate that I should raise my effective_cache_size on all the servers. My shared buffers settings are all 256 MB.
Does it matter that I have 2 instances running on each of these servers? IOW, should I divide the results in half, one for each instance and go with this as the recommended effective_cache_size setting instead?
Server 1 512 MB
Server 2 512 MB
Server 3 1 GB
Server 4 1 GB
Thanks for any insight,
Help yourself to FREE treats served up daily at the Messenger Café. Stop by today.
pgsql-admin by date
|Next:||From: Laszlo Nagy||Date: 2007-11-13 20:00:02|
|Subject: SHMMAX problem - how much is needed?|
|Previous:||From: Mike Blackwell||Date: 2007-11-13 19:22:33|
|Subject: Escape string tracing|