This is a simple example that demonstrates the use of transactions to protect consistency during concurrent access and modification of data. Since a region may configured across multiple servers and multiple clients may interact with that region independent of each other, data integrity relies on synchronization of modifications between all actors.

An example of how data can become inconsistent during concurrent interaction is as follows:

1. Client A gets the value of a key.
2. Client B gets the value of the same key.
3. Client A puts a new value for the key based upon the original value.
4. Client B puts a different new value for the key based upon the original value. The final value for that key is based upon the original value, not the updated value from Client A. For the final value to contain all the calculations from both clients, both the access and the modification of the value would need to happen as an atomic action across the region.

This example starts five child processes, each of which tries one thousand times to get the current value of a counter, increment that value, and the put the incremented value back into the region. To protect data consistency, the incrementing is abandoned and retried if another child has already incremented the value or if another child is simultaenously trying to increment the value. This example, which should take about a dozen seconds, reports the final value of the counter to show that all of the children's increments were consistently applied.

This example assumes you have installed Java and Geode.

1. From the geode-examples/transaction directory, build the example and run unit tests.

    $ ../gradlew build
   

2. Next start a locator, start a server, and create a region.

    $ gfsh run --file=scripts/start.gfsh
   

3. Run the example to demonstrate transactions.

    $ ../gradlew run
   

4. Shut down the system.

    $ gfsh run --file=scripts/stop.gfsh