README.md

Serverless Workflow Specification — Java SDK (Reference Implementation)- Persistence - MVStore

This document explains how to enable persistence using MVStore as underlying persistent mechanism. It is assumed that the reader is familiar with standard workflow execution mechanism.

To enable MVStore persistence, users should at least do the following things:

• Initialize a MVStorePersistenceStore instance, passing the path of the file containing the persisted information
• Pass this MVStorePersitenceStore as argument of DefaultPersistenceInstanceHandlers.from. This will create PersistenceInstanceWriter and PersistenceInstanceReader.
• Use the PersistenceInstanceWriter created in the previous step to decorate the existing WorkflowApplication builder.

The code will look like this

---

try (PersistenceInstanceHandlers handlers =
        DefaultPersistenceInstanceHandlers.from(new MVStorePersistenceStore("test.db"))
            .build();
    WorkflowApplication application =
        PersistenceApplicationBuilder.builder(
                WorkflowApplication.builder(),
                handlers.writer())
            .build(); )
 {
    // run workflow normally, the progress will be persisted
 }

---

If user wants to resume execution of all previously existing instances (typically after a server crash), he can use the reader created in the previous block to retrieve all stored instances.

Once retrieved, calling start method will resume the execution after the latest completed task before the running JVM was stopped.

---

  handlers.reader().scanAll(definition).forEach(WorkflowInstance::start);

---

---

Questions or ideas? PRs and issues welcome!