Skip to content

Latest commit

 

History

History
14 lines (12 loc) · 1.92 KB

service_quotas.md

File metadata and controls

14 lines (12 loc) · 1.92 KB

If you try to spin up a GPU and sometimes large CPU machines, you may encounter a quota limit. This is in place to prevent users from accidentally spinning up really expensive resources and burning through their whole budget in a few days.

If you would like to use a machine with a quota limit, you can follow these instructions.

Here are some Cloud Lab specific instructions as well.

  • Go to Services > Service Quotas
  • Click on Amazon Elastic Compute Cloud (EC2)
  • Search On-Demand, or whatever instance type you are trying to modify (e.g. Spot etc)
  • Click Running On-Demand P instances. You need to request whatever instance family you are trying to run. This can be figured out by visiting the EC2 instance type page and looking at the prefix of the instance type. For example, p4d.24xlarge will be from the P family so you need to request Running On-Demand P instances and so on for the other machine types.
  • Click Request quota increase
  • The quota value is the number of vCPUs allowed for an instance family, so set it to the max CPUs you expect to use
  • Click Request
  • In theory, you should only have to wait a few minutes for everything to be updated on the back end and then you can try and launch the blocked instance type again. In practice, it may take up to 30 minutes, so be patient and keep checking back until the Applied quota value is updated
  • If you are still having trouble, please contact STRIDES support