Rlabkey performance | Brian Connolly | 2012-08-14 18:49 |
Status: Closed | ||
Leo, If you were running just a LabKey Server on your workstation and not using it as a desktop too, I would recommend the following configuring the following setting - shared_buffers = 25% of system memory up to max of 2GB - work_mem = 10MB - maintenance_work_mem = 1024MB - effective_cache_size = 75% of system memory - wal_buffers = 20MB - checkpoint_segments = 15 - checkpoint_timeout = 15min - enable_material = off - random_page_cost = 1.5 - join_collapse_limit = 10 - track_counts = 1 - autovacuum = on Now for a workstation, like yours, where you will be doing many other things on it, you want to balance performance with resource usage. I recommend the following settings - shared_buffers = 512MB - work_mem = 10MB - maintenance_work_mem = 256MB - effective_cache_size = 75% of system memory - wal_buffers = 4MB - checkpoint_segments = 15 - [comment this out] checkpoint_timeout = 15min - enable_material = off - random_page_cost = 1.5 - join_collapse_limit = 10 - track_counts = 1 - autovacuum = on By default, most of these settings are commented out in the file postgresql.conf. To change the values, go to the file and uncomment the setting and change the value. Once you have made the changes, you will need to restart the postgresql server to pick up the changes. Also, if your workstation is running linux, you may need to set the kernel's Shared Memory setting. If you are on linux, and the postgresql server fails to start, after making these changes, then respond to this message and I will send you the instructions for increasing the Shared Memory settings on your work station. -Brian |
||