I have 2 Postgres 9.4 servers running on 2 identical VMs that I am using to learn how to setup replication before it gets used in production. Hot standby isn't required but it would be preferred. I setup the replication following this guide, though I'm on Ubuntu 14.04 using the Postgres repository, and quickly verified it works by running pgbench and looking at both servers.
Under the documentation for 9.4 for streaming replication, I need to set
wal_keep_segments to a high enough value in case the standby server gets behind, but does that only apply to asynchronous streaming replication, or does it also apply to synchronous? Forgive me if I missed that in the docs. Otherwise I'll have to setup replication slots to make sure the wal doesn't get behind.