Hi, I’m running into cycles quite often lately.
My main problem is that I use TF Cloud and I can’t even download the plan to get a graph and try to figure where the cycle is happening.
Does someone knows how to read the plan output in TF Cloud error for a cycle?
I need to learn if there’s a way to follow the close, expand, etc legends so I can figure what is causing the cycle.
There are no deposed resources involved.
I might send the ouput here but I would try to avoid it to keep the information as private as possible.
Thanks
Hi @notchjpl,
For the most part, you can try to follow just the base addresses of the nodes in the error output, ignoring the internal things like (expand) and (close). Normally walking through the cycle output (provided it’s not unreasonably long) along with inspecting the config will show where you have a block that ends up referring back to itself.
If you can post the cycle error output and configuration, someone can also take a look at it here. Make sure you are using a relatively up to date Terraform release too, as there were edge cases that were not handled correctly in early versions.
Thanks jbardin. I will try to follow your recommendation, can’t paste the config ATM but I have isolated the issue to the security groups. I use Cloudposse modules for ECS services and tasks and it looks like the dependency among the ALB and the service SG is the culprit of the cycle at the destroy time.
Thanks