From 6cfe341ee89baa952929e91d33b9ecbca73a3ea0 Mon Sep 17 00:00:00 2001 From: Shixiong Zhu Date: Tue, 5 Jan 2016 13:48:47 -0800 Subject: [SPARK-12511] [PYSPARK] [STREAMING] Make sure PythonDStream.registerSerializer is called only once There is an issue that Py4J's PythonProxyHandler.finalize blocks forever. (https://github.com/bartdag/py4j/pull/184) Py4j will create a PythonProxyHandler in Java for "transformer_serializer" when calling "registerSerializer". If we call "registerSerializer" twice, the second PythonProxyHandler will override the first one, then the first one will be GCed and trigger "PythonProxyHandler.finalize". To avoid that, we should not call"registerSerializer" more than once, so that "PythonProxyHandler" in Java side won't be GCed. Author: Shixiong Zhu Closes #10514 from zsxwing/SPARK-12511. --- python/pyspark/streaming/util.py | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) (limited to 'python/pyspark/streaming/util.py') diff --git a/python/pyspark/streaming/util.py b/python/pyspark/streaming/util.py index abbbf6eb93..e617fc9ce9 100644 --- a/python/pyspark/streaming/util.py +++ b/python/pyspark/streaming/util.py @@ -89,11 +89,10 @@ class TransformFunctionSerializer(object): it uses this class to invoke Python, which returns the serialized function as a byte array. """ - def __init__(self, ctx, serializer, gateway=None): + def init(self, ctx, serializer, gateway=None): self.ctx = ctx self.serializer = serializer self.gateway = gateway or self.ctx._gateway - self.gateway.jvm.PythonDStream.registerSerializer(self) self.failure = None def dumps(self, id): -- cgit v1.2.3