JVM crash before the updated value was sent out with gossip means that the update is "lost".
Use case for WriteOne is when you need low latency and are it's acceptable to loose updates if JVM crashes before the data has been spread to other nodes.
For example: web page counter
JVM crash before the updated value was sent out with gossip means that the update is "lost".
Use case for WriteOne is when you need low latency and are it's acceptable to loose updates if JVM crashes before the data has been spread to other nodes. For example: web page counter