Skip to content

genums/TPCCBench

 
 

Repository files navigation

------------------------------------------------------------------------------------------------
TPCC Benchmark Tools

This is a basic implementation of the TPC-C benchmark tool.

I use it for testing new database servers for throughput.
It also implements a heartbeat table for testing replication.
This doesn't generate certifiable results for TPC-C submission.

tpcc database schema.sql to generate the empty schema.
TPCCDatabaseGenerator.exe to initally populate the TPC-C database.
tpccbench.exe to execute the test.
------------------------------------------------------------------------------------------------
tpccbench.exe command line
------------------------------------------------------------------------------------------------
Commands are in \x:option format.
No spaces between the \x: and the option specified.
Valid command line arguements are:

\c:{1} number of clients you wish to run max 200

\s:{servername} name of the database server to benchmark

\d:{target database} the TPC database you wish to benchmark

\u:{username} login name to connect to server. For sql authentication only
              ignored if \t option is used

\p:{password} login password to connect to server. For sql authentication only
              ignored if \t option is used

\t use trusted connection notice no : or any following commands

\wh:{2} number of warehouses to use during tests.
         Depends on number of warehouses populated in test database.

\m:{1} number of minutes to execute the test minimum 1 no maximum

\cs:{50} maximum number of delay between queries issued by the Client.
         Any number between 0 and 50 is normal 0 for maximum throughput
         50 for a standard tpc benchmark Client thinking delay

\sp used extended stored procedures instead of single issued statements

\stl:{0} stagger Client load where {0} is the number of seconds to delay each Client

\nl:{0} number of fixed loops to execute per Client. This overrides the \m: switch

\hb Log to Heartbeat table to track replication latency.

Below is the work load mix switches all 5 must be used and must equal 100.
\pno:{45} percent new order status
\pos:{43} percent order status
\pp:{4} percent Payment
\pd:{4} percent dilivery
\psl:{4} percent stock level

\? for help

if there is a tpccbench.txt in the same directory it will read the command line switches from the file.
See the included file.

------------------------------------------------------------------------------------------------
execute tpcc database schema.sql to generate the schema
------------------------------------------------------------------------------------------------
TPCCDatabaseGenerator.exe command line
------------------------------------------------------------------------------------------------

Commands are in \Xoption format.
No spaces between the \X and the option specified.
A trusted connection is assumed.
Valid command line arguments are:
\SR Database Server Name
\D database name to load
\SW start warehouse ID Set this to 1
\EW end warehouse ID Set this to the upper limit i.e. 10 for 1 through 10 warehouses.
\TN table name to load Valid table names warehouse, item, district, customer, stock, order, neworder, orderline, history

see gendb.bat example for loading all tables in an empty database.

------------------------------------------------------------------------------------------------
I recommend backing up the database if you plan on running multiple tests. 
The restore will be quicker than regenerating the database every time.

About

A TPC-C style benchmarking tool written in c#

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • C# 100.0%