SwingBench can now be started to produce a transaction load and simulate an order-entry system to test the replication.
Start a new terminal session on the source server, as oracle. |
[oracle@source ~]$ cd $HOME/replicate [oracle@source replicate]$ ./run_swingbench.bash |
The output will be similar to the following and will run for 10 minutes. A total number of 10 users will be simulated.
Author : Dominic Giles Version : 2.5.0.932 Results will be written to results.xml. Time Users TPM TPS 8:39:02 PM 0 0 0 8:39:03 PM 0 0 0 8:39:04 PM 0 0 0 8:39:05 PM 2 0 0 8:39:06 PM 10 10 10 8:39:07 PM 10 23 13 <<output continues>> |
If you want to stop SwingBench before the 10 minutes is up, just hit Ctrl C in the terminal window where it's running to interrupt it. While the SwingBench is running - keep an eye on your dbvisit console (The dbvisit console should still be running. if not restart the dbvisit console on the source server - cd $HOME/dbvrep_XE and ./start-console.sh ) |
It is easy to change the parameters of swingbench to change the load and the time that it runs. The parameters are:
All command line options are listed here |
3. Optional. Swingbench is highly configurable. The load on the system can be easily increased. In this example, the number of users is set to 40 and the min and max delay is shortened. As a result the TPS peaks at over 400 transactions per second.
Your laptop has to be powerful enough to handle this load and a large number of archived redo logs will be generated that may fill up the disk space. |
/u01/app/oracle/swingbench/bin/charbench -cs //source-int/orcl -u oe -p oe -uc 40 -min 5 -max 15 -rt 0:10 -a -v users,tpm,tps Author : Dominic Giles Version : 2.5.0.932 Results will be written to results.xml. Time Users TPM TPS 20:53:05 [0/40] 0 0 20:53:06 [0/40] 0 0 20:53:07 [0/40] 0 0 20:53:08 [0/40] 0 0 20:53:10 [31/40] 0 0 20:53:13 [40/40] 440 440 20:53:14 [40/40] 478 38 20:53:18 [40/40] 913 435 20:53:19 [40/40] 967 54 20:53:22 [40/40] 1410 443 20:53:23 [40/40] 1446 36 20:53:25 [40/40] 1745 299 20:53:26 [40/40] 1885 140 |
Right, that should have given the replication something to do. Lets return to the Replication Console and see what's happening.