小编Asi*_*sim的帖子

我需要一些帮助来改善Cassandra的读取性能.随着列族的大小增加,我担心读取性能的下降.我们在单节点Cassandra上有以下统计数据.

操作系统: Linux - CentOS版本5.4(最终
版)Cassandra版本: apache-cassandra-1.1.0
Java版本: "1.6.0_14"Java(TM)SE运行时环境(版本1.6.0_14-b08)Java HotSpot(TM)64 -Bit Server VM(build 14.0-b16,混合模式)

Cassandra配置:(cassandra.yaml)

• rpc_server_type:hsha
• disk_access_mode:mmap
• concurrent_reads:64
• concurrent_writes:32

平台: Amazon-ec2/Rightscale m1.Xlarge实例,带有4个临时磁盘和raid0.(15 GB总内存,4个虚拟内核,2个ECU,总ECU = 8)

实验配置: 我尝试用GC做一些实验

Cassandra配置:
10 GB RAM分配给Cassandra Heap,3500MB是堆新大小.

JVM配置:
JVM_OPTS ="$ JVM_OPTS -XX:+ UseParNewGC"
JVM_OPTS ="$ JVM_OPTS -XX:+ UseConcMarkSweepGC"
JVM_OPTS ="$ JVM_OPTS -XX:+ CMSParallelRemarkEnabled"
JVM_OPTS ="$ JVM_OPTS -XX:SurvivorRatio = 1000"
JVM_OPTS = "$ JVM_OPTS -XX:MaxTenuringThreshold = 0"
JVM_OPTS ="$ JVM_OPTS -XX:CMSInitiatingOccupancyFraction = 40"
JVM_OPTS ="$ JVM_OPTS -XX:+ UseCMSInitiatingOccupancyOnly -XX:+ UseCompressedOops"

读取请求每秒208到240
写入请求每秒18到28个 …