hiro/ray
1
0
Fork 0
mirror of https://github.com/vale981/ray synced 2025-03-06 02:21:39 -05:00
Code Issues Projects Releases Packages Wiki Activity Actions

Set redis password in slurm deployment documentation. (#5747)

Browse source
This commit is contained in:
Robert Nishihara 2019-09-21 15:33:15 -07:00 • committed by GitHub
parent d17b35494d
commit c91a37f622
No known key found for this signature in database
GPG key ID: 4AEE18F83AFDEB23
1 changed files with 27 additions and 17 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

44
doc/source/deploying-on-slurm.rst
View file

@ -8,15 +8,15 @@ Clusters managed by Slurm may require that Ray is initialized as a part of the s
#!/bin/bash
#SBATCH --job-name=test
#SBATCH --cpus-per-task=20
#SBATCH --cpus-per-task=5
#SBATCH --mem-per-cpu=1GB
#SBATCH --nodes=5
#SBATCH --nodes=3
#SBATCH --tasks-per-node 1
worker_num=4 # Must be one less that the total number of nodes
worker_num=2 # Must be one less that the total number of nodes
module load Langs/Python/3.6.4 # This will vary depending on your environment
source venv/bin/activate
# module load Langs/Python/3.6.4 # This will vary depending on your environment
# source venv/bin/activate
nodes=$(scontrol show hostnames $SLURM_JOB_NODELIST) # Getting the node names
nodes_array=( $nodes )
@ -26,38 +26,48 @@ Clusters managed by Slurm may require that Ray is initialized as a part of the s
ip_prefix=$(srun --nodes=1 --ntasks=1 -w $node1 hostname --ip-address) # Making address
suffix=':6379'
ip_head=$ip_prefix$suffix
redis_password=$(uuidgen)
export ip_head # Exporting for latter access by trainer.py
srun --nodes=1 --ntasks=1 -w $node1 ray start --block --head --redis-port=6379 & # Starting the head
srun --nodes=1 --ntasks=1 -w $node1 ray start --block --head --redis-port=6379 --redis-password=$redis_password & # Starting the head
sleep 5
for (( i=1; i<=$worker_num; i++ ))
do
node2=${nodes_array[$i]}
srun --nodes=1 --ntasks=1 -w $node2 ray start --block --address=$ip_head & # Starting the workers
srun --nodes=1 --ntasks=1 -w $node2 ray start --block --address=$ip_head --redis-password=$redis_password & # Starting the workers
sleep 5
done
python trainer.py 100 # Pass the total number of allocated CPUs
python -u trainer.py $redis_password 15 # Pass the total number of allocated CPUs
.. code-block:: python
# trainer.py
from collections import Counter
import os
import sys
import time
import ray
ray.init(address=os.environ["ip_head"])
redis_password = sys.argv[1]
num_cpus = int(sys.argv[2])
ray.init(address=os.environ["ip_head"], redis_password=redis_password)
print("Nodes in the Ray cluster:")
print(ray.nodes())
@ray.remote
def f():
time.sleep(1)
time.sleep(1)
return ray.services.get_node_ip_address()
# The following takes one second (assuming that ray was able to access all of the allocated nodes).
start = time.time()
num_cpus = int(sys.argv[1])
ray.get([f.remote() for _ in range(num_cpus)])
end = time.time()
print(end - start)
for i in range(60):
start = time.time()
ip_addresses = ray.get([f.remote() for _ in range(num_cpus)])
print(Counter(ip_addresses))
end = time.time()
print(end - start)