You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Running l0-setup apply [instance] should complete without error.
Actual behavior
The end of the apply workflow incorporates the push workflow in order to backup the terraform files to S3. However, that apply doesn't seem to have access to the newly-created instance's credentials, so the push fails with AccessDenied. Using l0-setup endpoint to put the credentials into the environment and then re-running the same apply will then succeed, because apply will have access to those creds in the env vars.
Steps to reproduce the behavior
l0-setup apply [new-instance] and watch it fail at the end.
eval $( l0-setup endpoint -d -i [new-instance] | sed -n '1!p' ) to get env vars.
l0-setup apply [new-instance] and watch it succeed because of env vars.
Thoughts on how to proceed
The issue stems from an attempt to consolidate two commands into one without the prereq's for one of those commands. I have a couple of suggestions up front:
Fully separate the push workflow from the apply workflow. This makes the backing up of a layer0 instance's terraform files a responsibility of the user.
Incorporate the endpoint workflow into the apply workflow as well, so that when push is executed, it has the required credentials in the environment.
The text was updated successfully, but these errors were encountered:
Expected behavior
Running
l0-setup apply [instance]
should complete without error.Actual behavior
The end of the
apply
workflow incorporates thepush
workflow in order to backup the terraform files to S3. However, thatapply
doesn't seem to have access to the newly-created instance's credentials, so thepush
fails withAccessDenied
. Usingl0-setup endpoint
to put the credentials into the environment and then re-running the sameapply
will then succeed, becauseapply
will have access to those creds in the env vars.Steps to reproduce the behavior
l0-setup apply [new-instance]
and watch it fail at the end.eval $( l0-setup endpoint -d -i [new-instance] | sed -n '1!p' )
to get env vars.l0-setup apply [new-instance]
and watch it succeed because of env vars.Thoughts on how to proceed
The issue stems from an attempt to consolidate two commands into one without the prereq's for one of those commands. I have a couple of suggestions up front:
push
workflow from theapply
workflow. This makes the backing up of a layer0 instance's terraform files a responsibility of the user.endpoint
workflow into theapply
workflow as well, so that whenpush
is executed, it has the required credentials in the environment.The text was updated successfully, but these errors were encountered: