Bug 1962083

Description of problem:

Hello,

We must investigate and address https://github.com/kubernetes/kubernetes/issues/67761

Under specific circumstances (unknown to me), in namespaces with resourcequotas, Kubernetes will seemingly randomly return:
~~~
{"ErrStatus":{"kind":"Status","apiVersion":"v1","metadata":{},"status":"Failure","message":"Operation cannot be fulfilled on resourcequotas \"quota\": the object has been modified; please apply your changes to the latest version and try again","reason":"Conflict","details":{"name":"quota","kind":"resourcequotas"},"code":409}}
~~~

The issue happens once in a blue moon and can currently be reproduced with helm (our latest version and helm upstream).

For further details, please see: https://bugzilla.redhat.com/show_bug.cgi?id=1934604

See https://bugzilla.redhat.com/show_bug.cgi?id=1934604#c32 and https://bugzilla.redhat.com/show_bug.cgi?id=1934604#c33

I suggested retry logic for helm here: https://github.com/helm/helm/issues/9710

While retrying upon 409 Conflict errors should be good practice for clients, it should not be the task of every single client to catch the aforementioned resourcequota 409 as this looks like an internal issue to kubernetes / OpenShift.

If you look at https://github.com/kubernetes/kubernetes/issues/67761, a lot of clients are working around this currently, yet this kubernetes issue remained open for 2 years without a fix.

Something anecdotal: Our customer reported that this issue happens more often in their current version of OCP 4.6 (4.6.17) compared to the version of OCP which they ran before (4.5.16).


Version-Release number of selected component (if applicable):


How reproducible:


Steps to Reproduce:
1.
2.
3.

Actual results:


Expected results:


Additional info: