i want benchmark cassandra clusters 1, 2, 3 , 4 instances. ran cassandra-stress tool on one of nodes. benchmark shows strange results, see graph below (--> one-node cluster has more ops/sek 2-/3-/4-node cluster when few threads).
my results (x-axis=threads, y-axis=ops/sek, dataset=nodes in cluster (1, 2, 3, 4):
compared results this benchmark site, results seem not correct.
my question is: use tool correctly if run following command on 1 machine of cluster:
cassandra-stress write i tried without effect:
cassandra-stress write -node ip1,ip2,... see other question here. thank you!
-- edit: solution jim --
run cassandra-tool other ec2-instances outside c*-cluster, in same lan (so can work internal ips 10.x.x.x). launched 1/2/4 node cluster 4 separate benchmark-caller nodes. each of them got 1 of following commands:
first writing:
cassandra-stress write n=1000000 cl=one -mode native cql3 -schema keyspace="keyspace1" -pop seq=1..1000000 -node ip1,ip2,ip3,ip4 cassandra-stress write n=1000000 cl=one -mode native cql3 -schema keyspace="keyspace1" -pop seq=1000001..2000000 -node ip1,ip2,ip3,ip4 cassandra-stress write n=1000000 cl=one -mode native cql3 -schema keyspace="keyspace1" -pop seq=2000001..3000000 -node ip1,ip2,ip3,ip4 cassandra-stress write n=1000000 cl=one -mode native cql3 -schema keyspace="keyspace1" -pop seq=3000001..4000000 -node ip1,ip2,ip3,ip4 then reading data read command:
cassandra-stress read n=1000000 cl=one -mode native cql3 -schema keyspace="keyspace1" -pop seq=1..1000000 -node ip1,ip2,ip3,ip4 cassandra-stress read n=1000000 cl=one -mode native cql3 -schema keyspace="keyspace1" -pop seq=1000001..2000000 -node ip1,ip2,ip3,ip4 cassandra-stress read n=1000000 cl=one -mode native cql3 -schema keyspace="keyspace1" -pop seq=2000001..3000000 -node ip1,ip2,ip3,ip4 cassandra-stress read n=1000000 cl=one -mode native cql3 -schema keyspace="keyspace1" -pop seq=3000001..4000000 -node ip1,ip2,ip3,ip4 here results of reading
1 node cluster: 149,000 ops/sec 2 node cluster: 348,000 ops/sec 4 node cluster: 480,000 ops/sec
thank you, jim!
if running cassandra-stress on 1 node think expected result. single machine cannot saturate 4 node cluster , bottleneck.
also if running cassandra-stress on 1 of cassandra nodes, node doubly loaded running both cassandra , stress client. put strain on cpu , network connection machine.
to true picture of cluster throughput, should run stress multiple machines outside cluster (but on same lan).
Comments
Post a Comment