Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

SGE slots_per_host doesn't work #471

Closed
kevinbache opened this issue Nov 9, 2014 · 6 comments
Closed

SGE slots_per_host doesn't work #471

kevinbache opened this issue Nov 9, 2014 · 6 comments

Comments

@kevinbache
Copy link

Setting slots_per_host=1 doesn't seem to work. sshing in the master node and running qconf -mattr exechost complex_values slots=1 <hostname> does, however. Would you be interested in a pull request that changed the sge plugin to use this mechanism?

@cancan101
Copy link

What do you mean it doesn't work? I use this setting and it works as I expect

@kevinbache
Copy link
Author

I must be using some combination of settings that's invalidating it, then. With the config file below, the number of slots on each node remains at the default and they each suck up multiple jobs at once. Running qconf -se node001 on the master doesn't list any settings about slots either. I mentioned running qconf -mattr exechost complex_values slots=1 node001 from the master node because this added the slot setting to gconf and kept the worker nodes from consuming more than one job at once.

####################################
## StarCluster Configuration File ##
####################################
[global]
DEFAULT_TEMPLATE=kevin-template-west-1

[aws info]
AWS_ACCESS_KEY_ID = ######################
AWS_SECRET_ACCESS_KEY = ####################################
AWS_USER_ID= ####################
AWS_REGION_NAME = us-west-1
AWS_REGION_HOST = ec2.us-west-1.amazonaws.com

[key kevin-macbook]
KEY_LOCATION=~/.ssh/################

[vol kevin-vol-west-1]
volume_id = vol-#########
mount_path = /storage

[cluster kevin-template-west-1]
KEYNAME = kevin-macbook
CLUSTER_SIZE = 2
CLUSTER_USER = sgeadmin
CLUSTER_SHELL = bash
NODE_IMAGE_ID = ami-########
NODE_INSTANCE_TYPE = g2.2xlarge
DISABLE_QUEUE = True
PLUGINS = sge, git
MASTER_INSTANCE_TYPE = c3.large
SPOT_BID = 0.25
FORCE_SPOT_MASTER = True
VOLUMES = kevin-vol-west-1

[plugin sge]
setup_class = starcluster.plugins.sge.SGEPlugin
master_is_exec_host=False
slots_per_host=1

[plugin git]
setup_class = starcluster.plugins.git.GitCredentials
private_github_ssh_key_location = /Users/kevin/.ssh/########
git_user_name = Kevin Bache
git_user_email = [email protected]

@cancan101
Copy link

I have very similar settings and I do see that the number of slots is set to 1 on each of my hosts.
Why don't you add some debugging around:

num_slots = self.slots_per_host

also, I don't think you can change the config after the cluster has been started.

@kevinbache
Copy link
Author

Somehow this is working for me now. With apologies, I'm going to close this issue as irreproducible.

@cancan101
Copy link

What is the git plugin? If it is what I think it is, take a look at #468

@kevinbache
Copy link
Author

It's a little plugin I threw together to forward my git credentials to a cluster (ssh agent forwarding, set remote username/email) and pull some repos on cluster startup. The code is in my StarCluster fork. https://github.com/kevinbache/StarCluster/blob/develop/starcluster/plugins/git.py

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

2 participants