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.

Sign up for GitHub

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

Jump to bottom

Handle autoscaling of GPUs in GKE #392

Closed
vishh opened this issue Oct 13, 2017 · 1 comment
Closed

Handle autoscaling of GPUs in GKE #392

vishh opened this issue Oct 13, 2017 · 1 comment
Assignees
@bskiba
Labels
area/cluster-autoscaler

Comments

@vishh
Copy link

vishh commented Oct 13, 2017

Instance templates have accelerators included in them. The resource type expected to be used for GKE clusters to access GPUs is nvidia.com/gpus. Autoscalar should add to and remove from node-pools that have nvidia-tesla-[k80|p100] accelerator type included in them.

If the node pool containing GPUs is not the only node pool it is preferrable to scale all the way down to zero.
Output from gcloud instance-templates describe

guestAccelerators:
  - acceleratorCount: 8
    acceleratorType: nvidia-tesla-k80

cc @mwielgus
@bskiba bskiba assigned bskiba and unassigned mwielgus Oct 16, 2017
@bskiba bskiba mentioned this issue Oct 18, 2017
Merged
@bskiba
Copy link
Member

bskiba commented Oct 18, 2017

Fixed with #401

@bskiba bskiba closed this as completed Oct 18, 2017
@bskiba bskiba mentioned this issue Oct 19, 2017
Merged
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees

@bskiba bskiba

Labels
area/cluster-autoscaler
Projects
None yet
Milestone
No milestone
Development

No branches or pull requests
3 participants
@bskiba @vishh @mwielgus