Waltz is a distributed/replicated write ahead log for transactions.
We have the documention here.
To build Waltz:
./gradlew clean build
To run a single test class:
./gradlew clean :waltz-storage:test -Dtest.single=StorageClientTest
To run a single test method:
./gradlew clean :waltz-storage:test --tests com.wepay.waltz.storage.client.StorageClientTest.testBasicReadWrite
Waltz also comes with a smoke test that starts:
It then sends 1 million transactions while turning the servers on and off at random. Afterwards, it validates that all transactions were received, and that the checksums of the data are identical. It also logs throughput and latency.
To run smoke tests:
The smoke test output looks like:
0001[... .*. *] 0002[... .** *] 0003[... *** *] 0004[..* *** *] 0005[*.* *** *] 0006[*** *** *] --..-..-..-..-+..-..-+..-..-..-..-+..-..-..-..-..-..-+..+ 0007[.** *** *] +++-..-..-..- 0008[*** *** *] +..-..- 0009[*** .** *] 0010[*.* .** *] +++..-..-..-..-..-..+- 0011[*.* *** *] .. 0012[*** *** *] +-..-..-..-+..-+..-..-..-..+-..-
Each state change results in a new line. The first four digits indicate the number of server/storage start/stops that have been triggered so far. The
[... ... .] portion indicates the state of the server nodes, storage nodes and the zookeeper server, where
* means that the process is up, and
. means that it's down. Lastly, the trailing part of the line indicates 1000 writes (
-), 1000 reads (
.), and retries (
A log file (
smoketest.log) can be found in your home directory.
Waltz comes with a demo app that shows an example account balance database built on top of Waltz.
Start a test Waltz cluster in docker environment.
Start a MySQL instance in docker environment.
Run demo application with simple self-explanatory commands.
We implemented shell scripts to run a Waltz cluster in local Docker containers for testing purpose.
This builds the Docker images.
This creates a user defined docker network
starts three container, a zookeeper server, a waltz storage node, and a waltz server node in
Zookeeper port is 2181 inside
waltz-network and exposed to the host machine at 42181.
So, if you want to run a Waltz application outside of
The cluster's root ZNode is
/waltz_cluster. So specify this as
If the Docker images are not built yet, this script builds them.
However, it doesn't automatically build a new images even when the source code is modified.
You must rebuild images using
distDocker gradle task.
This stops all three containers. You can resume the cluster using
start command. All data in zookeeper and storages are preserved.
This will remove all three containers, thus removes all data.
First create a database. The following command will create a mysql container and the database.
Then, start the demo application:
To stop the MySQL instance:
To remove the database:
Go through website/README.md