From f4f9802c41828ff1da0c5bc9097f95ef23bc899c Mon Sep 17 00:00:00 2001
From: Sylvain Lebresne <sylvain@riptano.com>
Date: Tue, 29 May 2012 18:25:22 +0200
Subject: [PATCH] Make thrift HSHA thread limit unlimited by default

patch by scode; reviewed by slebresne for CASSANDRA-4277
---
 CHANGES.txt                                   |  1 +
 conf/cassandra.yaml                           | 37 ++++++++-----------
 .../cassandra/config/DatabaseDescriptor.java  |  9 ++---
 3 files changed, 20 insertions(+), 27 deletions(-)

diff --git a/CHANGES.txt b/CHANGES.txt
index 82ce164132..b36b736d7b 100644
--- a/CHANGES.txt
+++ b/CHANGES.txt
@@ -9,6 +9,7 @@
  * Add support for range tombstones (CASSANDRA-3708)
  * Improve MessagingService efficiency (CASSANDRA-3617)
  * Avoid ID conflicts from concurrent schema changes (CASSANDRA-3794)
+ * Set thrift HSHA server thread limit to unlimet by default (CASSANDRA-4277)
 
 
 1.1.1-dev
diff --git a/conf/cassandra.yaml b/conf/cassandra.yaml
index d64e2c9be0..eb01f38477 100644
--- a/conf/cassandra.yaml
+++ b/conf/cassandra.yaml
@@ -282,34 +282,29 @@ rpc_keepalive: true
 
 # Cassandra provides three options for the RPC Server:
 #
-# sync  -> One connection per thread in the rpc pool (see below).
-#          For a very large number of clients, memory will be your limiting
-#          factor; on a 64 bit JVM, 128KB is the minimum stack size per thread.
-#          Connection pooling is very, very strongly recommended.
+# sync  -> One thread per thrift connection. For a very large number of clients, memory
+#          will be your limiting factor. On a 64 bit JVM, 128KB is the minimum stack size
+#          per thread, and that will correspond to your use of virtual memory (but physical memory
+#          may be limited depending on use of stack space).
 #
-# async -> Nonblocking server implementation with one thread to serve 
-#          rpc connections.  This is not recommended for high throughput use
-#          cases. Async has been tested to be about 50% slower than sync
-#          or hsha and is deprecated: it will be removed in the next major release.
-#
-# hsha  -> Stands for "half synchronous, half asynchronous." The rpc thread pool 
-#          (see below) is used to manage requests, but the threads are multiplexed
-#          across the different clients.
+# hsha  -> Stands for "half synchronous, half asynchronous." All thrift clients are handled
+#          asynchronously using a small number of threads that does not vary with the amount
+#          of thrift clients (and thus scales well to many clients). The rpc requests are still
+#          synchronous (one thread per active request).
 #
 # The default is sync because on Windows hsha is about 30% slower.  On Linux,
 # sync/hsha performance is about the same, with hsha of course using less memory.
 rpc_server_type: sync
 
-# Uncomment rpc_min|max|thread to set request pool size.
-# You would primarily set max for the sync server to safeguard against
-# misbehaved clients; if you do hit the max, Cassandra will block until one
-# disconnects before accepting more.  The defaults for sync are min of 16 and max
-# unlimited.
-# 
-# For the Hsha server, the min and max both default to quadruple the number of
-# CPU cores.
+# Uncomment rpc_min|max_thread to set request pool size limits.
 #
-# This configuration is ignored by the async server.
+# Regardless of your choice of RPC server (see above), the number of maximum requests in the
+# RPC thread pool dictates how many concurrent requests are possible (but if you are using the sync
+# RPC server, it also dictates the number of clients that can be connected at all).
+#
+# The default is unlimited and thus provide no protection against clients overwhelming the server. You are
+# encouraged to set a maximum that makes sense for you in production, but do keep in mind that
+# rpc_max_threads represents the maximum number of client requests this server may execute concurrently.
 #
 # rpc_min_threads: 16
 # rpc_max_threads: 2048
diff --git a/src/java/org/apache/cassandra/config/DatabaseDescriptor.java b/src/java/org/apache/cassandra/config/DatabaseDescriptor.java
index 6d85425495..bb48ff76f3 100644
--- a/src/java/org/apache/cassandra/config/DatabaseDescriptor.java
+++ b/src/java/org/apache/cassandra/config/DatabaseDescriptor.java
@@ -381,13 +381,10 @@ public class DatabaseDescriptor
             if (!CassandraDaemon.rpc_server_types.contains(conf.rpc_server_type.toLowerCase()))
                 throw new ConfigurationException("Unknown rpc_server_type: " + conf.rpc_server_type);
             if (conf.rpc_min_threads == null)
-                conf.rpc_min_threads = conf.rpc_server_type.toLowerCase().equals("hsha")
-                                     ? Runtime.getRuntime().availableProcessors() * 4
-                                     : 16;
+                conf.rpc_min_threads = 16;
+
             if (conf.rpc_max_threads == null)
-                conf.rpc_max_threads = conf.rpc_server_type.toLowerCase().equals("hsha")
-                                     ? Runtime.getRuntime().availableProcessors() * 4
-                                     : Integer.MAX_VALUE;
+                conf.rpc_max_threads = Integer.MAX_VALUE;
 
             /* data file and commit log directories. they get created later, when they're needed. */
             if (conf.commitlog_directory != null && conf.data_file_directories != null && conf.saved_caches_directory != null)