We are facing problem related to performance of Postgres. Indexes are not being utilized and Postgres is giving priority to seq scan. I read many articles of Postgres performance and found that we need to set the randome_page_cost value same as seq_page_cost because we are using SSD drives. We are running copy of Discourse forum, you can read more about Discourse here meta.discourse.org. Details of all Server hardware and Postgres version are given below.
I am adding my Postgres configuration file in attachment, kindly review it and suggest the changes so that i can improve the performance of whole system. Currently queries are taking lot of time. I can also share the schema with you and queries in detail too.
Thanks
Postgres Version : 9.5.4
Server Hardware details : Dedicate machine 16 Physical cores 32 Logical cores RAM : 64 GB RAM Type : DDR3 Drive Type : SSD Raid Controller : MegaRAID SAS 2108 Raid Card Configured Raids : 10 No of Drives : 4 File System : XFS
