This sample demonstrates basic cluster performance allowing the user to vary the workload by specifying command line arguments as follows:
-w <thread_num> - the number of WRITE threads. More threads usually lead to higher overall performance
-l <obj_size> - the object size in bytes (min 8 bytes). Larger objects require more network bandwidth
-r <max_rate> - the maximum transaction rate for each thread (to simulate a partial load)
-n <trans_num> - the total number of transactions for each thread to process
-c <conflict_rate> - the maximum percent of conflicts
-u <update_rate> - the insert/update ratio, in percent (e.g. 50)
-t <max_time> - the maximum run time in seconds
-u <update_rate> - percent of updates
In a console window run:
cluster_demo 2 0Then in a second console window run:
cluster_demo 2 1But it is more instructive to run one instance in your development environment debugger, stepping through the code.