#terraform-aws-modules (2020-10)
Terraform Modules
Discussions related to https://github.com/terraform-aws-modules
Archive: https://archive.sweetops.com/terraform-aws-modules/
2020-10-05
Hi, does someone know the difference between:
• https://github.com/cloudposse/terraform-aws-eks-workers/
• https://github.com/cloudposse/terraform-aws-eks-node-group And if any, which one should I use? Deploying a new infra
Node Group is managed node group, where EKS controls eveything
eks-workers are unmanaged nodes (where you control everything)
eks-workers allow you more control and provide some features that managed node group does not have, e.g. it does not support Spot Instances
you also can use both if you have that use-case, in which case you connect both managed and unmanaged nodes to the same EKS cluster
Thank you for the explanation!
Do we have any guide / documentation to follow and create EKS cluster + NG or Fargate?
Not really
we have complete working and tested examples on EKS + NG https://github.com/cloudposse/terraform-aws-eks-node-group/blob/master/examples/complete/main.tf
I meant a guide or tutorial…
examples/complete is a fully working and tested solution.
there’s no tutorial/guide, but should just be to run terraform init , terraform apply
Thanks
2020-10-13
Hi all, I made a very smol PR to be reviewed: https://github.com/cloudposse/terraform-aws-ec2-instance-group/pull/16
We have forked the terraform-aws-ec2-instance-group to our own repo’s since we wanted to add multiple availability zone setup in the module, we have added this and now we are moving on to the next issue: There seems to be an issue with extra ebs volumes attached. While i was comparing our own module to yours I found this smol issue. Please let me know if there are any questions.
2020-10-14
Just a gentle nudge for https://github.com/cloudposse/terraform-aws-multi-az-subnets/pull/27 - I opened it back at the end of August. Tests have passed, needs an approval and merge, pelase!
@bazbremner Looks like you have merge conflicts sadly. Mind doing a rebase and we’ll get this merged?
Sure, just in the middle of a pairing session then I’ll rebase and push. I suspect it’ll need another approval after that push.
Sorry, took longer than I expected to get around to this. Rebased and pushed, waiting on the tests now.
Right, triggered the tests explicitly (nothing had happened after ~15 minutes - is that deliberate?), they’ve passed. @Matt Gowie over to you for another review and merge, please.
2020-10-15
I ran into a recent issue with: https://github.com/cloudposse/terraform-aws-eks-node-group concerning a cycle dependency. Has anyone else ran into this:
Error: Cycle: module.node-group.aws_eks_node_group.cbd[0] (destroy deposed 3fbf533a), module.node-group.aws_iam_role_policy_attachment.amazon_eks_worker_node_autoscaler_policy[0] (destroy), module.node-group.aws_eks_node_group.cbd[0]
?
@Jeremy G (Cloud Posse)
The current terraform-aws-eks-node-group requires Terraform v0.13.3 or later. This is documented in the release notes.
It was ran via atlantis using tfv0.13.4
@Cody Moore I suggest you double-check that, as this error is expected when you run TF 0.13.2 or earlier and was fixed in TF 0.13.3
Hmm interesting ok
Will do, thank you
Here are the release notes: https://github.com/cloudposse/terraform-aws-eks-node-group/releases/tag/0.12.0
Yep definitely running terraform 0.13.4
Cycle is in different order if that makes any difference:
Error: Cycle: module.node-group.aws_iam_role_policy_attachment.amazon_eks_worker_node_autoscaler_policy[0] (destroy), module.node-group.aws_eks_node_group.cbd[0], module.node-group.aws_eks_node_group.cbd[0] (destroy deposed 39e8a12f)
Yea, still stumped on this.. Tabling for now and will try again later. Any advice for debugging the cycle?
No, the order does not matter, nor do the start and end points (it is a cycle, after all). I suggest trying with TF 0.13.3. We have not used TF 0.13.4 with this module and it is possible there was a regression in Terraform. Also try TF 0.14. If it still does not work, I suggest looking at this TF bug and seeing if it suggests anything to you with respect to fixing the issue or opening a new bug report.
i got such issue when using terraform terraform_0.12.24 after updating aws-modules to latest one
I ended up manually deleting one of the resources then reapplying, that worked
@Jeremy G (Cloud Posse) has joined the channel
2020-10-19
What’s the state of https://github.com/cloudposse/terraform-aws-cloudtrail-cloudwatch-alarms ? I noticed that the repo is pretty far behind now with TF 0.13 out. I was going to help with the update but realized that it was partially done in a few PRs
@Cody Moore It’s common that there are some older modules that just haven’t been updated to 0.12 syntax because CP started saying “We’re only going to upgrade modules to 0.12 if they have tests and pass those tests as that is the only sustainable way that we can support 200+ modules”.
So it looks like there were attempts to upgrade to 0.12, but none of them went all the way through the process. If you really want to bring that module into the new world then please feel free to fork the furthest along of those PRs, update it as far as you can to add tests and get them passing, and put it up on PR. I’d be happy to review and help you along in the process.
Couple example PRs of what is required from a 0.11 => 0.12 upgrade —
Gotcha, makes sense. Updating to 0.12 is really the first goal then right? With that said, the main testing infra used is CodeFresh like in the example PR given: https://github.com/cloudposse/terraform-aws-cloudfront-s3-cdn/pull/45/files ?
Also, a lot of the PRs are close as well, so I can also just use them as reference, since there isn’t just one that is the ‘furthest’ along imo
We’re no longer using CodeFresh. Check out the iam-user module PR above. CodeFresh was removed from the CP module toolset in between 0.12 and 0.13. We’re now using GH actions and there is an associated “ChatOps” workflow file for that.
And yeah, if you want start fresh and use the existing PRs against that module as reference then go for it.
Ah yep, ok I see that now. Thanks for the clarification. I’ll take a stab at that. Thanks!
No problem! Let me know if you have any other questions while you’re going through it.
Will do
2020-10-26
Hi all,
I’ve been working with the EKS terraform modules, and I ran into an issue with scaling nodegroups from this repo - https://github.com/cloudposse/terraform-aws-eks-node-group
So the problem is that I try to increase desired_size by specifying higher value, however the changes for desired_size are being ignored because of the following code in the [main.tf](http://main.tf)
lifecycle {
create_before_destroy = false
ignore_changes = [scaling_config[0].desired_size]
}
Can anyone explain why desired_size has to be ignored in this situation ?
2020-10-31
Does anyone know why the module bitflight-public/terraform-aws-app-mesh which looks like a CloudPosse module is no longer available via the CloudPosse GitHub Org?