-
Notifications
You must be signed in to change notification settings - Fork 74
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Failed to check app setting propagation for SCM_DO_BUILD_DURING_DEPLOYMENT and ENABLE_ORYX_BUILD. #86
Comments
Seeing the same issue with a .NET Core 3.1 Function deployment. Wasn't seeing this a couple of weeks ago. |
Also seeing the same issues on .NET 5.0 Function deploy. I attempted most of the changes you did, I even changed those settings in Azure to see if it would read them from the function and skip setting them but no luck. It's adding a ton of time to our deploys. |
Same here with a Python using this template and a brand new Azure Functions with Python runtime 3.8 (although other versions behave similar). It seems due to recent changes as reverting back to v.1.3.1 fixes it: - name: 'Run Azure Functions Action'
uses: Azure/[email protected] |
This issue is marked need-to-triage for generating issues report. |
+1 - I can repro if I create the pipeline from the Azure Portal directly |
+1 - Also experiencing the same issue. Function deployment duration has slowed from 1 minute to 8 - 10 minutes. |
+1 - We are also experiencing this issue, switching back to v1.3.2 fixes it. We are using the action to deploy a Python 3.9 project, |
This seems like a backend issue where the Kudu lite container is not able to pick up the app settings change in time. |
Not to sound negative bc I appreciate the issue getting attention but that doesn't sounds like it's fixing the actual problem. And all of us who are having the issue will still see increased build times (just not quite as bad, bc of 3 vs 10 retries). Speaking only for myself, I would like to see the core issue get resolved or to have these attempts to apply settings removed altogether if they're not actually needed. |
Reducing the retry count will treat the symptom (delay in function app deployment) but not the root cause (settings not syncing with kudu container). It is never successfully syncing these settings with the Kudu container, unless I revert to v1.3.2 as suggested above. Thank you for looking into this! |
This is a bandage solution to alleviate the symptoms of the AppSettings not propogating to Kudu as mentioned in Azure#86 . Because the propagation is broken both these settings retry 20 times at a 5 second interval and 200 seconds (over 3 minutes) are added to any builds using this action because of it. Setting the retry count still allows 15 seconds of redundancy but wastes far less time (just 30s) if the setting just fails to copy. Ultimately, I feel the settings propagation should either be fixed, or this setting removed if the propagation is unnecessary, but I don't know enough about those settings to understand how to fix it or if they are required.
I opened #90 as a bandage solution to limit the retries while a proper solution is figured out. |
Can confirm that I am experiencing this issue as well. Azure Functions 3.1 new GitHub workflow. |
- avoid a problem relates to this issue Azure/functions-action#86
* Add or update the App Service deployment workflow configuration from Azure Portal. * Specify the version of `Azure/functions-action` - avoid a problem relates to this issue Azure/functions-action#86 * Separate slot by target branch * Update code
This is what I get as of today. It looks like a complete failure of the deployment process. |
Reverting to v1.3.2 fixed it for me. |
Reverting to v1.3.2 fixed it for me as well. Was using v1.0.0 and v1.4.0. |
Reverting to v1.3.2 of the action fixed the error for me also. I can provide debug logs if needed. |
Should this issue be "idle"? I think we're just waiting around to see it get fixed. |
Yes, a lot of time and no response whatsoever and it is even properly identified by the repo maintainers. This is a bug that considerably slows downs release times on our Prod/Dev instances for each Azure Function that we have, and it is a lot of time to be wasting on a CI/CD pipeline. |
Just discovered this today too. Deploying Azure Functions. 19 retries and then succeeds. Slows down deploy time dramatically. |
Also getting the same - action completes successfully, however spends ages retrying over the two variables. |
Same here:
after that:
and finally:
|
On my case, I do the builds using ubuntu-latest.
3 days ago my deployments stopped working. The github actions logs were saying that the deployment was succesful, but my functions were NOT deployed.
Rolling back to v1.3.2 fixed the issue to me. I assume that my case was being a failure because I need these 2 settings to be true since I'm building on ubuntu instead of windows.
|
Hello folks - thanks for reporting it. we are looking into this. |
This bug is fixed in the latest version v1.4.2 (also v1). |
@patelchandni While no longer getting retry messages about the settings, I have yet to have a deploy complete successfully since trying the updates (previously pinned at My latest attempt has spent over 18 minutes on the Just like before, reverting to |
Just tried. Same results as @mikesprague. 18 minutes for me with @v1 and @v.1.4.2. |
@mikesprague @morrisonbrett |
@balag0 Unless you require option 2, I manually ran one of my functions in the function app and got the following: GUID: Full output:
|
You can see the action runs here: https://github.com/morrisonbrett/TrueVote.Api/actions You can also see in the commit history where I used |
The hanging deployment issue should be fixed in the latest v1.4.3 (v1). Please give it a try and let us know. |
Confirmed fixed. I just pushed using |
@patelchandni Working for me now as well with the latest (via |
Worked as well for me when changing to v1. Do you guys recommend keeping v1 in my deployment scripts or locking it to 1.4.3? |
@zhughes3 It is recommended to use v1 to get the latest version. |
As a general rule, using a v1 tag should be ok - though we've seen an example in this action where there was a regression so using v1 meant you got hit with the bug. An alternative is to use a specific version and enable Dependabot to provide PRs to upgrade your actions. That way you're more intentional about changing things. |
This issue is idle because it has been open for 14 days with no activity. |
FYI for apps with flex consumption see #245 (comment) |
First observed yesterday, our deployment action is having some issues. We're using node.js, our action is based on this template.
After
Client.updateAppSettingViaKudu
is called and returns 204, we get stuck in a timeout / retry cycle for about 3 minutes. See Github job log below.Here's some additional information.
Here's a few things I've tried, to no effect.
scm-do-build-during-deployment
andenable-oryx-build
in the workfile (both true & false).NODE_VERSION
from 12.x to 10.xI'd be happy to provide more information on request. Thanks for your help!
The text was updated successfully, but these errors were encountered: