Slow docker image download from GCP
Hi, I am experimenting with runpod recently. I tried to deploy a whisper image to runpod from my company GCP docker repo and I found it pretty slow. It look almost 10 minutes to download a 11GB size image. While I understand the image is huge, but I wonder is ther any things to do to speed up the process. For example, the repo location (current in asia, as my company is in asia)
7 Replies
maybe closer to the host ( us, eur )
@KAL how many layers?
i resolved by moving to US region and removing poetry cache, which slim by docker image from 11 to 5
fyi @Madiator2011 (Work) 19 layers
btw is there any way to tigger a new release with API
In template, if you update image version it will trigger a new release
https://docs.runpod.io/sdks/graphql/manage-pod-templates for anyone who search in discord
Manage Templates | RunPod Documentation
Create, modify, and delete templates in RunPod using GraphQL API with various parameters for container disk size, Docker arguments, environment variables, and more.
i just tried to call saveTemplate to trigger a new release, but it keep returning
INTERNAL_SERVER_ERROR
. Can you help me check what is the problem? thanks
Also it will be nice to just requiring id
and things I want to change only for saveTemplate
, currently i need to refill env and else which is not going to change for a new releaes 🤔
id, imagename and name are changed in discord
Try to write your feedback in #🧐|feedback
Yeah idk which di required the error is like not really clear