Millicores and OutOfSync

[timmc-edx]

I'm using ArgoCD v2.8.0+804d4b8 (yes, I know that's old) and have encountered what seems like an ArgoCD bug around resource limits.

I increased a container's CPU limits and requests to 1.0 and 0.25 respectively, but sync never finished. ArgoCD shows the below manifest diff, where the current limit is '1' and the desired limit is 1, and the current request is 250m and the desired request is 0.25.

You'll notice that these actually mean the same thing, and the sync should have completed.

• Is this an ArgoCD bug or a Kubernetes bug?
• Can anyone else reproduce this on a newer ArgoCD?
• Is this fixed in a newer ArgoCD?

[rumstead]

#3357

Your version should have the fix but I suggest upgrading.

[timmc-edx]

Ah, thanks for the pointer! Yeah, I wonder if there was a regression that has also since been fixed. I'll see if we can upgrade... :-)