I have been monitoring resource utilization on node managers through Ambari metrics. Cache memory is keep increasing on all node manager nodes.
Does it impact performance?
our cluster has been installed on Centos 7.3.
Can you please clarify my doubt?
Thanks in Advance.
Typically cached memory would be evicted when it is needed. Is this OS cache or HDFS cache ? you can read more about centralized cache memory mgmt in HDFS here -