This project is a work in progress.
Please see our Google Group for discussion.
cassandra-stress is a configuration-based tool for doing benchmarks and testing simple datamodels for Apache Cassandra. Unfortunately it can be difficult to configure a workload. There are faily common data models and workloads seen on Apache Cassandra, this tool aims to provide a means of executing configurable, pre-defined profiles.
Full docs are here (sort of, work in progress): http://thelastpickle.com/tlp-stress/
Clone this repo, then build with gradle:
git clone https://github.com/thelastpickle/tlp-stress.git
cd tlp-stress
./gradlew assemble
Use the shell script wrapper to start and get help:
bin/tlp-stress -h
Time series workload with a billion operations:
bin/tlp-stress run BasicTimeSeries -i 1B
Key value workload with a million operations across 5k partitions, 50:50 read:write ratio:
bin/tlp-stress run KeyValue -i 1M -p 5k -r .5
Time series workload, using TWCS:
bin/tlp-stress run BasicTimeSeries -i 10M --compaction "{'class':'TimeWindowCompactionStrategy', 'compaction_window_size': 1, 'compaction_window_unit': 'DAYS'}"
Time series workload with a run lasting 1h and 30mins:
bin/tlp-stress run BasicTimeSeries -d "1h 30m"