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
terraform plan should ensure, as much as technically possible, that apply will perform successfully. Meaning that when plan is ok, apply should be ok. This is clearly not the case today.
Actual Behavior
plan step is not helping enough. Thus, when using terraform enterprise, one has to merge his feature branch to master to troubleshoot. A few examples of things that went wrong during apply, that were not detected during plan when creating an akamai_property:
Static validations:
Invalid property name
one of product,product_id must be specified
invalid client access token / secret / ... because of a trailing \n: should have been caught earlier than during the API call, where the message is not helping (Invalid header...)
Checks against Akamai:
Property name already in use
product_access_denied
Product not found
The text was updated successfully, but these errors were encountered:
Thanks for submitting the issue. I have a couple of questions and comments.
First. of all, I agree with the part about static validations, that is absolutely something that can be improved and we will work on applying your suggestions for the next release. Validating edgerc credentials format can be added, as well as checking whether either product or product_id is specified during planning phase. Can you however clarify what do you mean by checking whether property name is valid? I'm not sure what kind of validation can we add there, as none are specified by the open API documentation.
As to checking against Akamai - unfortunately, if state is not available, terraform does not invoke any methods from terraform provider when preparing the plan - all it does is compare your config with resource's schema (that's where static validation takes place), but we are not able to interact with the API at that point. That is how other terraform provider work as well, e.g. AWS.
Hi there,
Expected Behavior
terraform plan should ensure, as much as technically possible, that apply will perform successfully. Meaning that when plan is ok, apply should be ok. This is clearly not the case today.
Actual Behavior
plan step is not helping enough. Thus, when using terraform enterprise, one has to merge his feature branch to master to troubleshoot. A few examples of things that went wrong during apply, that were not detected during plan when creating an akamai_property:
The text was updated successfully, but these errors were encountered: