We're still considering this architecture - so I don't have any plans.
My question is, does the ANALYZE command, which is only executed on the master, mean that the statistics / plans that are used on the master are ALSO used on the slaves? OR .... does the slave create it's own plan?
I would anticipate very different plans between the two machines, and different postgresql.conf settings as well.
I would *want* different query plans between the two. Is that possible?