Gruntwork
Upgrading terraform-aws-lambda past version v0.21.0 #685
-
|
I'm trying to upgrade our Lambda edge code that uses terraform-aws-lambda to v0.21.0 but I'm having issues with the new requirements for However, when I do that I get errors deploying stating that Here is the full output of the error. Previously I had my own log group that was created. I removed that with Terraform even though it had a different name. I looked in modules/lambda-edge-multi-region-log-groups/main.tf and it looks like all of the log groups have the same name. If I comment out all but one of the log group regions it still fails. Below is my module code. This ends up with this Terraform plan: But I still get the following errors: |
Beta Was this translation helpful? Give feedback.
Replies: 5 comments 5 replies
-
|
Thanks for reporting this issue, @erictompkins. I'll have a look at this, do some investigations, and come back to you in a bit :) I was the engineer who worked on this feature but this was a while ago and I need to refresh my memory! Thanks for the patience |
Beta Was this translation helpful? Give feedback.
-
|
So, to explain a bit of what's going on: The log groups can have the same name in different regions, that's not an issue. The problem is the following: if you have lambda edge, these log groups are created in every region by AWS automatically. So they end up escaping terraform control. Because all of the resources created by the modules should be managed by terraform, this is why this feature was added. So now these log groups are created explicitly and can be controlled by terraform and be deleted properly or receive custom configuration through terraform, etc. But they were created anyway before by AWS when you used edge, this is why you are having the "already exists" error. Now, nothing of what I said so far actually helps you solve it 😅 So I'm gonna continue to have a look at this, see how we can help you do this migration, and get back to you soon. Thanks! Edit to add this: you should not try removing the log regions, unless it's a region you have completely disabled on your AWS account. That's because, as I said, these log groups are created either way. |
Beta Was this translation helpful? Give feedback.
-
|
@Etiene Thank you for looking into this. We tried to update an existing Lambda Edge function and got this issue. We also tried to create a new Lambda Edge function and also got this issue. |
Beta Was this translation helpful? Give feedback.
-
As I said, that's expected, the solution is probably related to using
This, however, is not expected! Because an issue like this should have been caught by our tests. So I'm going to have to investigate this separately. Did you destroy the previous function and tried to create a new one? Did it have the same name or was this with a completely separate function? |
Beta Was this translation helpful? Give feedback.
-
|
I’m pretty sure that i destroyed the previous function and then tried to
add a new one with the same name.
I have to create some new ones in a different account so I’ll try again.
On Tue, Mar 21, 2023 at 1:33 PM Etiene Dalcol ***@***.***> wrote:
We tried to update an existing Lambda Edge function and got this issue.
As I said, that's expected, the solution is probably related to using terraform
import to add these groups to your state, so I was looking into this to
create a better migration guide for that release.
We also tried to create a new Lambda Edge function and also got this issue.
This, however, is not expected! Because an issue like this should have
been caught by our tests. So I'm going to have to investigate this
separately. Did you destroy the previous function and tried to create a new
one? Did it have the same name or was this with a completely separate
function?
—
Reply to this email directly, view it on GitHub
<#685 (reply in thread)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/ABFAM4GUEON62AZM3KQPO3TW5HRALANCNFSM6AAAAAAV7COFJM>
.
You are receiving this because you were mentioned.Message ID:
***@***.***
com>
--
Eric Tompkins
President / Founder - Aptuitiv, Inc
Celebrating 16 years of service!
Professional Website Design & Development
Aptuitiv, Inc
207-860-4040
PO Box 129
Farmington, ME 04938
www.aptuitiv.com
|
Beta Was this translation helpful? Give feedback.
-
|
I see! I think that what happened was, because you were on a version where these groups escaped terraform, they remained there after destruction (the exact problem this feature wants to avoid in the future). So when you created a function with the same name, there they were. But thanks for testing it again with new names, while I keep looking into this here. Let me know if that also does not work. |
Beta Was this translation helpful? Give feedback.
-
|
I created a brand new Lambda function using All of the Cloudwatch groups were created successfully. Your explanation of what's happening with an existing Lambda function makes sense. I'm not sure how to update an existing function. It sounds like importing with I appreciate your help. |
Beta Was this translation helpful? Give feedback.
-
|
Hi @erictompkins, great to hear that it worked for new functions! I'm currently on holidays and am back on April 10th. If you're not in a rush I can help you then, otherwise I can ask one of my colleagues to support you. Thanks for the patience :) |
Beta Was this translation helpful? Give feedback.
-
|
@Etiene April 10th is ok. Enjoy your vacation! Thank you |
Beta Was this translation helpful? Give feedback.
-
|
Hi @erictompkins, could you try this and let me know if it will apply without errors? Replace YOUR_FUNCTION_NAME below. You might also have to replace |
Beta Was this translation helpful? Give feedback.
Hi @erictompkins, could you try this and let me know if it will apply without errors?
Replace YOUR_FUNCTION_NAME below. You might also have to replace
module.lambda_edgeaccording to how your module is named. If you have more "log group already exists" errors, you can follow this pattern to import the remaining log groups according to the error message you get.