Cassandra performance slow down with counter column
- by tubcvt
I have a cluster (4 node ) and a node have 16 core and 24 gb ram:
192.168.23.114 datacenter1 rack1 Up Normal 44.48 GB 25.00%
192.168.23.115 datacenter1 rack1 Up Normal 44.51 GB 25.00%
192.168.23.116 datacenter1 rack1 Up Normal 44.51 GB 25.00%
192.168.23.117 datacenter1 rack1 Up Normal 44.51 GB 25.00%
We use about 10 column family (counter column) to make some system statistic report.
Problem on here is that When i set replication_factor of this keyspace from 1 to 2 (contain 10 counter column family ), all cpu of node increase from 10% ( when use replication factor=1) to --- 90%. :( :(
who can help me work around that :( . why counter column consume too much cpu time :(.
thanks all