A.Shuffle is not a necessary process in Spark tasks

B.Spark can cache intermediate results in memory reducing disk I0

C.Intermediate results of Spark tasks can only be stored in memory

D.Shuffle is a necessary process in Spark tasks

A.When the number of files in WALs reaches a threshold

B.The total size of the MemStore in the Region has reached the preset Flush Size threshold

C.HBasePeriod refreshMemstore, silentthink period is 1Hour

D.The ratio of the total memory occupied by the MemStore to the total memory of the RegionServer exceeds the preset threshold size

A.The amount of leaf swallowing of the entire distributed file system is limited by the throughput of a single name node

B.The number of objects (files, blocks) that a name node can hold is limited by the size of the memory space

C.Failure of this unique name node will render the entire cluster unavailable

D.It will affect the functional implementation of the upper-level components based on HDFS

A.Support mixed deployment of Kunpeng server and ordinary x86 server in a single cluster

B.No shortcomings in performance

C.Achieve autonomous controllability of some equipment

D.Single component (such as HDFS) supports mixed deployment of Kunpeng server and x86 server

A.False

B.True