Chapter 47. Conflict Manager Usage

The Conflict Manager is often used with Partition Handling. A split-brain occurs when nodes in a cluster are separated into two or more groups (partitions) that can’t communicate with each other. In some split-brain situations, nodes can have different data written to them. In this case, JBoss Data Grid’s Partition Handling, combined with its Conflict Manager, can be used to automatically resolve differences in the same CacheEntries across nodes. The Conflict Manager can also be used to manually search for and resolve conflicts.

The code below shows how to retrieve an EmbeddedCacheManager’s ConflictManager, how to retrieve all versions of a given key, and how to check for conflicts across a given cache.

EmbeddedCacheManager manager = new DefaultCacheManager("example-config.xml");
Cache<Integer, String> cache = manager.getCache("testCache");
ConflictManager<Integer, String> crm = ConflictManagerFactory.get(cache.getAdvancedCache());

// Get All Versions of Key
Map<Address, InternalCacheValue<String>> versions = crm.getAllVersions(1);

// Process conflicts stream and perform some operation on the cache
Stream<Map<Address, InternalCacheEntry<Integer, String>>> stream = crm.getConflicts();
stream.forEach(map -> {
   CacheEntry<Object, Object> entry = map.values().iterator().next();
   Object conflictKey = entry.getKey();
   cache.remove(conflictKey);
});

// Detect and then resolve conflicts using the configured EntryMergePolicy
crm.resolveConflicts();

// Detect and then resolve conflicts using the passed EntryMergePolicy instance
crm.resolveConflicts((preferredEntry, otherEntries) -> preferredEntry);