Hi Dave,
It is much easier to backup/snapshot the entire VRMS appliance in case you are using embedded DB. This is a preferable way to have a protection over your VRMS configuration. In case you are using external DB you have to make snapshots of the VRMS appliance and the external DB VM. It's a good idea to make those snapshots approximate at the same time.
In case you still want to try to manually backup your VRMS DB you should know that it is not so easy and also it is not supported in any way. Not only embedded VRMS DB should be backed up but also embedded VR DB - located in a file something like /etc/vmware/hbrsrv.xx.db. Also you should back up file /etc/vmware/hbrsrv.uuid.conf and maybe certificates files also. VRMS DB can be dumped with /opt/vmware/vpostgres/1.0/bin/pg_dumpall > /tmp/dbdump -Uvrmsdb .
Restoring VR DB is just a file replacement (you should stop VRMS and VR services before that). Restoring VRMS DB is done by dumping the current one with /opt/vmware/vpostgres/1.0/bin/psql postgres -c "drop database vrmsdb" -Uvrmsdb and importing from backed up file with command : /opt/vmware/vpostgres/1.0/bin/psql -f /tmp/dbdump postgres -Uvrmsdb . In case you have deployed new VRMS appliance and you're trying to configure it from back ups you will have to manually edit data in the DB because the new VRMS appliance VM has a new MOID so the one stored in the DB should be replaced.