Hey guys We’re trying to implement the metaflow-ui module from terraform-aws-metaflow. It almost succeeds in creating the ECS cluster: two out of three services fail deploying: • ui_backend ❌ • ui_static ✅ • metadata_service ❌ both errors are the same:

CannotPullContainerError: ref pull has been retried 1 time(s): failed to copy: httpReadSeeker: failed open: unexpected status code <https://registry-1.docker.io/v2/netflixoss/metaflow_metadata_service/manifests/sha256:a3ce2a20f32ab37537fd67003afac1b99b165ba8c89f561ce73ce78f6ba6c081>

do you know why it can’t pull that container?

I think I ran into a similar problem when redeploying our trial stack to our main region. I found a tread about ECR images and updated to point to them instead of dockerhub

@clean-wire-89462 what is your input value for:

metadata_service_container_image

when running terraform? I see that the digest in your error message is for

v2.3.0

and I am able to pull that image from docker hub. The same tag exists in public ECR as well, just with a different digest.

thanks @jolly-afternoon-77975, I’ve managed to fix it by using the containers from ECR! however, the

ui_alb_dns_name

was not outputted from the ui module and the dns I defined is not working, even after adding my ip to the allow list. do you have any insight?

Great to hear. Unfortunately not, i had to get my admin to setup the UI and dont have an insight. But I dont think we ran into ui_alb_dns_name error.

that’s strange, did you have to define all these required variables? https://github.com/outerbounds/terraform-aws-metaflow/tree/master/modules/ui

I am using the cloud formation template for it

@clean-wire-89462 are you adding your IP to the ALB's security group? I did run into a similar issue where calls to the UI were just hanging because of a security group rule and had to expand the inbound rule to be able to hit the UI. p.s. I am new at outerbounds and am currently working through the same terraform code to understand the infra needed to run metaflow and also look for any gaps or issues with tf/cf code and accompanying documentation. Any feedback you have will be greatly appreciated.

I might know a guy who did it recently 🙂 @ripe-car-38698 he’s based in EST time zone, so later. But afaik he deployed the UI recently

Hey guys, anyone interested in a kick debug with me? everything was implemented according to doc but still does not work

Hey @clean-wire-89462 are you available now to just on a call?