Github user StefanRRichter commented on a diff in the pull request: https://github.com/apache/flink/pull/5518#discussion_r169069584 --- Diff: flink-contrib/flink-statebackend-rocksdb/src/main/java/org/apache/flink/contrib/streaming/state/RocksDBKeyedStateBackend.java --- @@ -266,9 +267,16 @@ public RocksDBKeyedStateBackend( RocksIterator iterator = db.newIterator(columnInfo.f0); iterator.seekToFirst(); - Iterable<K> iterable = () -> new RocksIteratorWrapper<>(iterator, state, keySerializer, keyGroupPrefixBytes); - Stream<K> targetStream = StreamSupport.stream(iterable.spliterator(), false); - return targetStream.onClose(iterator::close); + try { + ByteArrayOutputStream outputStream = new ByteArrayOutputStream(8); + namespaceSerializer.serialize(namespace, new DataOutputViewStreamWrapper(outputStream)); + final byte[] namespaceBytes = outputStream.toByteArray(); + Iterable<K> iterable = () -> new RocksIteratorWrapper<>(iterator, state, keySerializer, keyGroupPrefixBytes, namespaceBytes); + Stream<K> targetStream = StreamSupport.stream(iterable.spliterator(), false); + return targetStream.onClose(iterator::close); + } catch (IOException ex) { + throw new FlinkRuntimeException("Failed to get keys from RocksDB state backend.", ex); --- End diff -- This case would not call `iterator.close()`. I suggest to just create everything related to `namespaceBytes` before creating the iterator, so that the iterator is only created when no further exception should happen.
---