Package org.infinispan.conflict

Interface EntryMergePolicy<K,​V>

    • Method Summary

      All Methods Instance Methods Abstract Methods 
      Modifier and Type Method Description
      org.infinispan.container.entries.CacheEntry<K,​V> merge​(org.infinispan.container.entries.CacheEntry<K,​V> preferredEntry, List<org.infinispan.container.entries.CacheEntry<K,​V>> otherEntries)
      This method is called by ConflictManager.resolveConflicts() for each conflict discovered to determine which CacheEntry should be utilised.

    • Method Detail

      • merge

        org.infinispan.container.entries.CacheEntry<K,​V> merge​(org.infinispan.container.entries.CacheEntry<K,​V> preferredEntry,
                                                             List<org.infinispan.container.entries.CacheEntry<K,​V>> otherEntries)
        This method is called by ConflictManager.resolveConflicts() for each conflict discovered to determine which CacheEntry should be utilised. This merge policy is used when a user explicitly calls ConflictManager.resolveConflicts() as well as when a partition merge occurs with PartitionHandling.ALLOW_READ_WRITES set. In the event of a partition merge, we define the preferred partition as the partition whom's coordinator is coordinating the current merge.
        Parameters:
        preferredEntry - During a partition merge, the preferredEntry is the primary replica of a CacheEntry stored in the partition that contains the most nodes or if partitions are equal the one with the largest topologyId. In the event of overlapping partitions, i.e. a node A is present in the topology of both partitions {A}, {A,B,C}, we pick {A} as the preferred partition as it will have the higher topologId because the other partition's topology is behind. During a non-merge call to ConflictManager.resolveConflicts(), the preferredEntry is simply the primary owner of an entry
        otherEntries - a List of all other CacheEntry associated with a given Key.
        Returns:
        the winning CacheEntry to be utilised across the cluster, or null if all entries for a key should be removed.