Hello, trying to follow the deployment guide "AWS ...
# ask-metaflow
a
Hello, trying to follow the deployment guide "AWS Managed with CloudFormation" on the Outerbounds site; I have tried deploying the cloud formation template as stock as possible on a new AWS account a few times; and keep getting the same error `2024/08/30 044344 goose run: pq: no pg_hba.conf entry for host "10.20.0.94", user "master", database "metaflow", no encryption`; the ECSFargateService resource in the stack is stuck on CREATE_IN_PROGRESS and eventually times out and fails. I have left everything on default, tried with both the additional configuration fields changed or left as default, tried changing the deployed instance size (thinking the account may be prevented from deploying certain instance sizes as a new account); and not had any luck. Does anybody have any ideas of anything else to try? Thank you.
👀 1
✅ 1
b
can you try out the changes in this PR to see if they fix the issue: https://github.com/outerbounds/metaflow-tools/pull/58
this is a recent change if I recall,
rds.force_ssl
is enabled on engine versions 15 and up, but was disabled until v14
a
Thank you very much for the response, I tried the file in that PR and am still getting the same error; I have created a pastebin of the full logs if that is useful https://pastebin.com/qjFmyMMh Any other ideas?
b
oh, the other thing I noticed is that the image pinned in the cloudformation is quite old. specifically regarding providing ssl_mode for the connection, the stable version that has that is
2.4.6
I believe.
I'd suggest going with the latest version of the service unless there are any specific reasons to try older ones.
updated my branch with the latest pinned version. Hope this solves the issue 🙂
✅ 1
a
Thank you for the help, it your changes in the PR solve the issue.
b
awesome, glad to hear 👌