Closed aterrel closed 1 year ago
$ aws --version
aws-cli/2.11.11 Python/3.11.3 Darwin/22.5.0 source/arm64 prompt/off
Hm, haven't seen this before but AWS libraries often give us new fun edge cases.
If you feel like hacking things, I suspect that changing line 1100 of
/Users/aterrel/miniconda3/envs/storyfit_models_pipeline/lib/python3.10/site-packages/coiled/cli/setup/aws.py
to
quota_client = session.client("service-quotas", region_name=region)
would fix this (and I have PR to make this change).
Thanks for the quick response. Sam got the install to work with the command line flag, I'm happy to test your PR but not sure the steps to do so.
Glad you got something working!
If you don't mind testing the change I made, you could pip install coiled==0.8.9.dev6
and run coiled setup aws --quotas
and see if you get an error.
Seems to work for me.
$ coiled setup aws --quotas
╭──────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────╮
│ Coiled runs in your AWS account, so your cluster sizes will be constrained by your AWS quotas: │
│ │
│ Standard On-Demand 4512 vCPU │
│ Standard Spot 1152 vCPU │
│ G4dn (NVIDIA T4 GPU) On-Demand 920 vCPU │
│ G4dn (NVIDIA T4 GPU) Spot 64 vCPU │
│ P (NVIDIA V100/A100 GPU) On-Demand 692 vCPU │
│ P (NVIDIA V100/A100 GPU) Spot 64 vCPU │
│ │
│ Standard includes general purpose M and T families (e.g., M6i, T3), compute optimized C families (e.g., C6i), and memory optimized R families │
│ (e.g., R6i). │
│ │
│ GPU instances require a separate quota. G4dn is our default GPU instance type on AWS, these have an NVIDIA T4 GPU. │
│ │
│ The default Coiled instance type (t3.xlarge) has 4 vCPUs, so a 10 worker cluster (plus scheduler) would have 44 vCPUs. │
╰──────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────╯
Your current quota for Standard On-Demand (us-east-1) is 4512. If you'd like to request an increase from AWS, enter the new number (or hit return to
skip): (4512):
Your current quota for Standard Spot (us-east-1) is 1152. If you'd like to request an increase from AWS, enter the new number (or hit return to
skip): (1152):
Your current quota for G4dn (NVIDIA T4 GPU) On-Demand (us-east-1) is 920. If you'd like to request an increase from AWS, enter the new number (or
hit return to skip): (920):
Your current quota for G4dn (NVIDIA T4 GPU) Spot (us-east-1) is 64. If you'd like to request an increase from AWS, enter the new number (or hit
return to skip): (64):
Your current quota for P (NVIDIA V100/A100 GPU) On-Demand (us-east-1) is 692. If you'd like to request an increase from AWS, enter the new number
(or hit return to skip): (692):
Your current quota for P (NVIDIA V100/A100 GPU) Spot (us-east-1) is 64. If you'd like to request an increase from AWS, enter the new number (or hit
@aterrel thanks for confirming that the fix worked for you!
(this is matt from andy's computer)
Config for profile: