I have a Directory server with a read-only replication partner. The servers and all my client workstations are set to use TLS for commuication, and the clients are set up with with both hostnames in /etc/ldap.conf.

During a recent maintenace period, we noticed that if we take down the primary server, it takes 5 minutes for the clients to realize the primary is down, and to switch over to the backup sever. I'm assuming this is due to the 300 second timeout for the TLS session. Is there a way to shorten this to a more acceptable time?