-
Notifications
You must be signed in to change notification settings - Fork 3k
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
Internal Server Error on what-if analysis after upgrading to 2.28.0 from 2.26.0 #19850
Comments
Thanks for the feedback! We are routing this to the appropriate team for follow-up. cc @armleads-azure. Issue Details
Describe the bug Reverting to AZ 2.26.0 results in the what-if and the deployment both working. See the full What-If debug log here: https://github.com/Azure/Aks-Construction/runs/3863296553?check_suite_focus=true#step:11:146
To Reproduce az deployment group what-if --debug -f bicep/main.bicep -g $RG -p .github/workflows_dep/AksDeploy-ByoVnet.parameters.json -p resourceName=$RESNAME byoAKSSubnetId=*** byoAGWSubnetId=*** Expected behavior Environment summary Additional context
|
route to service team |
@shenglol Could you please help to have a look at this issue? |
I can confirm this is also an issue on 2.29.0 |
I'm still trying to figure out where the internal error comes from. It doesn't seem to be returned by the What-If service. |
Found the root cause. This is indeed a bug in the What-If engine. We added normalization process for Azure KeyVault access policies in 2021-01-01, but there's a case we failed to handle. I'm going to fix it. |
Hi @shenglol - my workaround of using an older version of the Az CLI is broken because of Azure/cli#56 - So i'm forced to use v2.30.0 of the Az CLI until either
Do you have a view of what CLI version your fix will make it into? |
This is a bug in our service with API version 2021-01-01 and after. We have checked in a fix and it should be rolled out in a about 2 weeks. Once it's rolled out any CLI version should work. |
Hi, any update on the rollout progress ? I'm getting the same error with az version:
Is there any workaround I can apply while fix is rolling e.g. switch specific bicep resources to a different api version or something else ? Interesting, though
Any hints for workaround are greatly appreciated. |
Additional note: when I run
UPD: Getting some inconsistent results - sometimes --verbose helps, sometimes it doesn't. Feels like a subtle race condition... |
@centur This is most likely because the requests were picked up by worker jobs in different regions. Up to this point, the fix is rolled out to 5 regions, and it still needs more time to be fully deployed. |
@shenglol Can I somehow affect what workers pick up my jobs ? We have resources in AU East and Southeast regions and it's affecting our productivity a lot - deployments are failing randomly and there is nothing we can do to work around it. Upd: also prob ignore my guesses about how to get bicep code working with --debug or --verbose flags - it's ended up being a mere co-incidence that it was working for me on the day of testing and after that I'm getting all kind of random results. Those flags are more of a cargo cult practice I inferred when tried to find some predictability in results on my end. They don't work the way I thought they were |
I'm now seeing
On previously working actions that were using 2.30.0 with no template/parameter changes |
@Gordonby I just realized you might be using the MSFT tenant which is onboarded to deployments preview features...and this appears to be a new bug we just identified in the recently added preview feature that enables The bug is basically an unhandled edge where a null ref exception will be thrown if the referenced resource in the template does not contain the The current workaround would be to replace all user assigned identity property accesses in
We have committed a fix for this, but given the upcoming holiday deployment freeze, it might take an extended time for the fix to be rolled out. My apologies for any inconvenience caused! |
@centur There's no way to control that. Do you mind sharing your ARM template and emailing me it at [email protected]? I am curious to see if I can provide a workaround, but I won't be able to tell without seeing the contents of the template. |
Yes - i'm using the Microsoft tenant. Thanks for the workaround note. |
I got the exact same error when running with what-if but fine without it. Weird thing is this only happens in one subscription but not in the other. The error is below. My az cli version is 2.32.0 and bicep version is 0.4.1318. |
@harrchen88 Could you share your Bicep template? |
Describe the bug
Performing a what-if analysis results in an internal server error, i'm unable to decipher the problem. The actual deployment then continues to works fine on 2.28.0.
Reverting to AZ 2.26.0 results in the what-if and the deployment both working.
See the full What-If debug log here: https://github.com/Azure/Aks-Construction/runs/3863296553?check_suite_focus=true#step:11:146
To Reproduce
Use the bicep files and the parameter file.
Run the following command, substituting two subnet resource id's and an arbitary resource name.
Expected behavior
I'd like a clear error message of what's actually failed, or for it just to work as it did in the previous CLI version.
Environment summary
GitHub Az CLI Action.
Additional context
https://github.com/Azure/Aks-Construction/runs/3863296553?check_suite_focus=true#step:11:146
The text was updated successfully, but these errors were encountered: