I am trying to set up a OLTP like workload, giving the example like this:
10% 8K read hit
40% 8K read miss
30% 8K write miss
10% 64K seq read
10% 64K seq write
I am using "skew" to distribute the workload, Is the following WD definition will do the work I expect to do ?
- wd=wd1,sd=sd*,xfersize=8k,rdpct=100,rhpct=100,seekpct=100,skew=10
- wd=wd2,sd=sd*,xfersize=8k,rdpct=100,rhpct=0,seekpct=100,skew=40
- wd=wd3,sd=sd*,xfersize=8k,rdpct=0,whpct=0,seekpct=100,skew=30
- wd=wd4,sd=sd*,xfersize=64k,rdpct=100,seekpct=0,skew=10
- wd=wd5,sd=sd*,xfersize=64k,rdpct=0,seekpct=0,skew=10
rd=OLTP,wd=(wd1,wd2,wd3,wd4,wd5),curve=(10,20,30,40,50,60,70,80,90,95,99),iorate=curve,interval=1,warmup=20,elapsed=300,pause=0,forthread=1