Github user redsanket commented on a diff in the pull request:
https://github.com/apache/spark/pull/22173#discussion_r218942035
--- Diff:
common/network-common/src/main/java/org/apache/spark/network/TransportContext.java
---
@@ -77,17 +82,54 @@
private static final MessageEncoder ENCODER = MessageEncoder.INSTANCE;
private static final MessageDecoder DECODER = MessageDecoder.INSTANCE;
+ // Separate thread pool for handling ChunkFetchRequest. This helps to
enable throttling
+ // max number of TransportServer worker threads that are blocked on
writing response
+ // of ChunkFetchRequest message back to the client via the underlying
channel.
+ private static EventLoopGroup chunkFetchWorkers;
+
public TransportContext(TransportConf conf, RpcHandler rpcHandler) {
- this(conf, rpcHandler, false);
+ this(conf, rpcHandler, false, false);
}
public TransportContext(
TransportConf conf,
RpcHandler rpcHandler,
boolean closeIdleConnections) {
+ this(conf, rpcHandler, closeIdleConnections, false);
+ }
+
+ /**
+ *
+ * @param conf TransportConf
+ * @param rpcHandler RpcHandler responsible for handling requests and
responses.
+ * @param closeIdleConnections Close idle connections if it is set to
true.
+ * @param isClientOnly This config is more important when external
shuffle is enabled.
+ * It stops creating extra event loop and
subsequent thread pool
+ * for shuffle clients to handle chunked fetch
requests.
+ * In the case when external shuffle is disabled,
the executors are both
+ * client and server so both share the same event
loop which is trivial.
--- End diff --
I hope we follow a similar indentation for all other comments
---
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org
For additional commands, e-mail: reviews-h...@spark.apache.org
