Sameer, It is mainly to guard for concurrent access for interactive queries:
https://kafka.apache.org/0110/documentation/streams/developer-guide#streams_interactive_queries In Kafka Streams, we allow users to independently query the running state stores in real-time in their own caller thread while the application's thread is continuously updating these stores, and hence the synchronization. Guozhang On Sun, Aug 13, 2017 at 10:24 PM, Sameer Kumar <sam.kum.w...@gmail.com> wrote: > Hi, > > I was inspecting RocksDBStore and I saw get and put methods being > synchronized. > > Are there any issues in RocksDB due to which we have used it. Could someone > please enlighten more on this. > > @Override > public synchronized V get(K key) { > validateStoreOpen(); > byte[] byteValue = getInternal(serdes.rawKey(key)); > if (byteValue == null) { > return null; > } else { > return serdes.valueFrom(byteValue); > } > } > > @SuppressWarnings("unchecked") > @Override > public synchronized void put(K key, V value) { > validateStoreOpen(); > byte[] rawKey = serdes.rawKey(key); > byte[] rawValue = serdes.rawValue(value); > putInternal(rawKey, rawValue); > } > > -Sameer. > -- -- Guozhang
