Now think of is, will the unstructured data timing is copied to the flume to monitor folder, collection, there will be no other production environment design,

CodePudding user response:

Before me from all kinds of application system of the location of the log server log directory to get to them, thinking the same with you, have a monitored directory, copy regularly, to install the flume on several system node, the data mining to the later, to kafka, spark and the background according to the topic of kafka consumption,