Hi, I wanted to understand more on how a broker recovers from a failure. As far as I understand, when a broker recovers from a failure, it 1) loads data of all partitions it hosted from the disk (the metadata about the partitions would exist in zookeeper, right?)2) then syncs with the leader of all the leader partitions to sync any data that was written during the recovery period
And along with that it might recover other metadata too. Is this understanding correct? If yes, does it make sense to go with a network storage for each broker in a private/public cloud environment? Because the broker would want to recover most of the data from the disk rather than over the network. Any documentation or blogs around this would help. Thanks,Chirag