Improvements to documentation and error messages. (#221)

doc/install-on-macosx.md

@@ -10,7 +10,6 @@ To install Ray, first install the following dependencies. We recommend using
 ```
 brew update
 brew install cmake automake autoconf libtool boost wget
-sudo easy_install pip  # If you're using Anaconda, then this is unnecessary.
 
 pip install numpy cloudpickle funcsigs colorama psutil redis --ignore-installed six
 ```
@@ -21,7 +20,7 @@ Ray can be built from the repository as follows.
 
 ```
 git clone https://github.com/ray-project/ray.git
-cd ray/lib/python
+cd ray/python
 python setup.py install --user
 ```
 

doc/install-on-ubuntu.md

@@ -10,7 +10,7 @@ To install Ray, first install the following dependencies. We recommend using
 
 ```
 sudo apt-get update
-sudo apt-get install -y cmake build-essential autoconf curl libtool python-dev python-pip libboost-all-dev unzip  # If you're using Anaconda, then python-dev and python-pip are unnecessary.
+sudo apt-get install -y cmake build-essential autoconf curl libtool libboost-all-dev unzip python-dev python-pip  # If you're using Anaconda, then python-dev and python-pip are unnecessary.
 
 pip install numpy cloudpickle funcsigs colorama psutil redis
 ```
@@ -21,7 +21,7 @@ Ray can be built from the repository as follows.
 
 ```
 git clone https://github.com/ray-project/ray.git
-cd ray/lib/python
+cd ray/python
 python setup.py install --user
 ```
 

doc/using-ray-on-a-cluster.md

@@ -27,10 +27,14 @@ should look something like `123.45.67.89:12345`).
 ```
 ./ray/scripts/start_ray.sh --redis-address <redis-address>
 ```
-To specify the number of processes to start, use the flag `--num-workers`, as follows:
+
+To specify the number of processes to start, use the flag `--num-workers`, as
+follows:
+
 ```
 ./ray/scripts/start_ray.sh --num-workers <int>
 ```
+
 Now we've started all of the Ray processes on each node Ray. This includes
 
 - Some worker processes on each machine.
@@ -39,7 +43,8 @@ Now we've started all of the Ray processes on each node Ray. This includes
 - One Redis server (on the head node).
 - One global scheduler (on the head node).
 
-To run some commands, start up Python on one of the nodes in the cluster, and do the following.
+To run some commands, start up Python on one of the nodes in the cluster, and do
+the following.
 
 ```python
 import ray
@@ -56,7 +61,7 @@ def f(x):
 ray.get([f.remote(f.remote(f.remote(0))) for _ in range(1000)])
 ```
 
-### Stopping Ray 
+### Stopping Ray
 When you want to stop the Ray processes, run `./ray/scripts/stop_ray.sh`
 on each node.
 

python/ray/services.py

@@ -114,10 +114,7 @@ def cleanup():
       successfully_shut_down = successfully_shut_down and success
     # Reset the list of processes of this type.
     all_processes[process_type] = []
-  if successfully_shut_down:
-    if len(all_processes) > 0:
-      print("Successfully shut down Ray.")
-  else:
+  if not successfully_shut_down:
     print("Ray did not shut down properly.")
 
 def all_processes_alive(exclude=[]):

scripts/stop_ray.sh

@@ -1,3 +1,9 @@
 #!/usr/bin/env bash
 
-killall redis-server global_scheduler plasma_store plasma_manager photon_scheduler
+killall global_scheduler plasma_store plasma_manager photon_scheduler
+
+# Find the PID of the Redis process and kill it.
+kill $(ps aux | grep redis-server | awk '{ print $2 }') 2> /dev/null
+
+# Find the PIDs of the worker processes and kill them.
+kill $(ps aux | grep default_worker.py | awk '{ print $2 }') 2> /dev/null

test/runtest.py

@@ -123,7 +123,7 @@ class CustomError(Exception):
   import cloudpickle
   cloudpickle.dumps(Point)
 except AttributeError:
-  cloudpickle_command = "pip install --upgrade git+git://github.com/cloudpipe/cloudpickle.git@0d225a4695f1f65ae1cbb2e0bbc145e10167cce4"
+  cloudpickle_command = "pip install --upgrade cloudpickle"
   raise Exception("You have an older version of cloudpickle that is not able to serialize namedtuples. Try running \n\n{}\n\n".format(cloudpickle_command))
 
 class SerializationTest(unittest.TestCase):