Cool. How much ram do you have on your server and what is the size of the database?
My settings of 32 GB was set due to recommendation of 25% from here [1]. Right now I am going to investigate the content of buffer cache with pg_buffercache and maybe OS page cache with recently released pg_stat_kcache to determine the optimal size for my load profile.
I have read several topics [1, 2, 3, 4] with similar problems but haven't find a good solution. Below is some more diagnostics.
I reported the second one. The diagnostics was very similar to yours. I think a lot people experienced this problem with big servers. It was only because of large shared_buffers. The problem disappeared after we reduced it to 2 GB.
