Ukri Niemimuukko

The usage of hard-coded IP 10.54.34.56 requires a bit of explaining. Perhaps the folder name "archives" could be reconsidered also. My initial thinking would have been that perhaps some old...

> @ttzeng IMHO examples in this project should: > * Be self-contained and simple enough that users can run them _as-is_ > => I think more complex examples should be...

> In case of `RemovePodsViolatingTopologySpreadConstraint`, setting `nodeFit` to true or false will yield the same result. Which is confusing and should be (at minimum) documented On this, I agree wholeheartedly....

Nit for self, and for conversation: PodFitsAnyNode has no users anymore if this gets merged. Should it be removed at the same time?

Squashed and cleaned. Sorry for the slipped in hacks in the intermediates.

> @uniemimu thanks for keeping the defaulting in mind. As part of moving to the descheduling framework, this becomes part of the core changes in the code base after all...

@RyanDevlin I verified that your patches fix this

/remove-lifecycle stale

Since @RyanDevlin is busy and #636 isn't moving overly fast, may I propose an alternative approach for making sure this nearly 6 months old bug gets fixed. 1) I push...

@damemi I took a look at the legacy here: https://github.com/kubernetes-sigs/descheduler/commit/40bb490f4cd2f5b1ee218e03c25c6df223ae8313#diff-92b554150104f06be4ac18898266bad157dc2f611a6268d873765bf9608aadb3R56-R60 And indeed you are correct in that the legacy is such that it fits better `NodeFit` being `true` considering how...