Configuration records for heat-config were stored under /var/run, which were lost on node reboot. The cause for this was reported and fix (see BZ#1278181) but had an impact on the OSP director; Puppet reapplied old manifests during an Overcloud update, which caused various Pacemaker cluster errors. The fix for BZ#1278181 moves the configuration records from /var/run/heat-config to /var/lib/heat-config. This fix also includes the "heat-config-rebuild-deployed" script to rebuild the configuration records. Now updates are possible without reapplying old Puppet manifests. Make sure to follow the documented update procedure, including running the "heat-config-rebuild-deployed" script on each node.