When an agent commits its CDB data, CDB will stream the committed data out to all registered slaves. If a slave dies during the commit, nothing will happen, the commit will succeed anyway. When and if the slave reconnects to the cluster, the slave will have to copy the entire configuration.
This “copy the entire configuration” by slave side is auto-performed by slave ConfD itself, or is required to be performed by HAFW, namely a process other than ConfD?
The CDB replication process is taken care of by ConfD. As long as ConfD is told by the HAFW to become a slave, it will contact the master and start the copying as necessary. The job of the HAFW is as described in Chapter 23.2, HA framework requirements, of the ConfD User Guide:
ConfD only replicates the CDB data. ConfD must be told by the HAFW which node should be master and which nodes should be slaves.
The HA framework must also detect when nodes fail and instruct ConfD accordingly. If the master node fails, the HAFW must elect one of the remaining slaves and appoint it the new master. The remaining slaves must also be informed by the HAFW about the new master situation. ConfD will never take any actions regarding master/slave-ness by itself.