search

marqo-ai / marqo

Unified embedding generation and search engine. Also available on cloud - cloud.marqo.ai
https://www.marqo.ai/
Apache License 2.0
4.51k stars 185 forks source link

[BUG] #820

Closed Arseniy-II closed 5 months ago

Arseniy-II commented 5 months ago

Describe the bug Docker images freeze on Amazon Linux 2023 AMI (with 30 Gb of space) for marqo:2.0.0 and marqo:latest

To Reproduce Steps to reproduce the behaviour:

  1. Go to AWS
  2. Lunch EC2 t2 micro instance with Amazon Linux 2023 AMI
  3. SSH into instance
  4. Install docker and try to run marqo:2.0.0 or marqo:latest

Expected behaviour Vector database up and running

Actual result Stoped at step "Marqo is deploying the application and waiting for the response from document API to start..." and nothing is happening for 3 hours

Screenshots If applicable, add screenshots to help explain your problem.

Desktop (please complete the following information):

Additional context Screenshot 2024-04-24 at 14 41 29 Screenshot 2024-04-24 at 14 42 00

tomhamer commented 5 months ago

Hey Arseniy, thanks for reaching out. The absolute minimum amount of RAM you can use to run marqo is 8GB. We have a ticket out so that marqo will have a more helpful error in this case: https://github.com/marqo-ai/marqo/issues/796

The least expensive instance you can use to run marqo is an r7g.medium, but I would recommend a g4dn.xlarge for the best performance. If you want to get this into production, the marqo kubernetes apps or marqo cloud are the best options.

https://github.com/marqo-ai/marqo-on-kubernetes

Arseniy-II commented 5 months ago

@tomhamer Thanks! I was able to lunch it without any issues on r7g.medium!